Syntax in Functional Grammar

Syntax in Functional Grammar

This page intentionally left blank

Syntax in Functional Grammar An introduction to lexicogrammar in systemic linguistics

G. DAVID MORLEY

CONTINUUM London and New York

Continuum Wellington House 125 Strand London WC2R OBB

370 Lexington Avenue New York NY 10017-6503

First published 2000 © G. David Morley 2000 All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage or retrieval system, without prior permission in writing from the publishers. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN 0-8264-4734-1 (hardback) 0-8264-4735-X (paperback) Library of Congress Cataloging-in-Publication Data Morley, G. David (George David), 1943Syntax in functional grammar : an introduction to lexicogrammar in systemic linguistics / by G. David Morley. p. cm. Includes bibliographical references and index. ISBN 0-8264-^734-1 - ISBN 0-8264-4735-X (pbk.) 1. Grammar, Comparative and general - Syntax. 2. Systemic grammar. I. Title. P295.M67 2000 415-dc21 00-021925 Typeset by Paston PrePress Ltd, Beccles, Suffolk Printed and bound in Great Britain by Biddies Ltd, Guildford and King's Lynn

Contents

Introduction

1

Part I The linguistic framework 1

The systemic functional framework 1.1 The linguistic system 1.2 Context of situation 1.3 Language functions and the semantic stratum 1.4 The systemic orientation

7 7 8 11 17

2

The grammatical framework 2.1 The nature of the lexicogrammar 2.2 Units and the rank scale 2.3 Sentence, clause, phrase, word and morpheme 2.4 Unit complex and complex unit

21 21 23 25 27

Part II

Class

3

Word class 3.1 Noun 3.2 Verb 3.3 Adjective 3.4 Adverb 3.5 Preposition 3.6 Conjunction 3.7 Interjection 3.8 Article 3.9 Pronoun 3.10 Determiner: a functional element 3.11 Labelling the word classes

31 31 33 39 40 42 43 45 45 45 50 51

4

Phrase class 4.1 Nominal phrase 4.2 Verbal phrase 4.3 Adjectival phrase 4.4 Adverbial phrase 4.5 Prepositional phrase 4.6 Subordinator phrase 4.7 Genitive phrase 4.8 Labelling the phrase classes

53 53 55 55 55 56 56 57 58

vi

Contents

5

Clause class 5. 1 Main and subordinate clauses 5.2 Nominal clause 5.3 Adjectival clause 5.4 Adverbial clause 5.5 Labelling the clause classes

59 59 63 64 66 68

6

Sentence class 6. 1 Formal types of sentence 6.2 The sentence as clause complex 6.3 Labelling the sentence

69 69 70 71

7

Formal syntactic analysis

72

Part III

Structure

8

Sentence structure 8. 1 Free and bound elements 8.2 Traditional classification of bound elements 8.3 Contemporary classification of free and bound elements 8.4 Integral and supplementary bound elements 8.5 Functional analysis of sentence structure

9

Clause structure 9.1 Subject 9.2 Predicator and Finite 9.3 Complement 9.4 Adjunct 9.5 Z element 9.6 Functional analysis of clause structure

91 91 95 99 104 108 109

10

Elements of clause structure revised 10.1 Determining the elements of clause structure 10.2 Nominally functioning elements 10.2.1 Subject 10.2.2 Object 10.2.3 Complement 10.3 Adjectivally functioning elements 10.4 Adverbially functioning elements 10.4.1 Circumstantial adverbials 10.4.2 Adjunct adverbials 10.5 Predicator 10.6 Functional analysis of clause structure revised

111 111 113 113 114 116 117 117 117 120 122 123

11

Phrase structure 11.1 Nominal phrase 11.1.1 Determiner 11.1.2 Modifier 11.1.3 Qualifier 1 1 .2 Adjectival phrase 11.3 Adverbial phrase 1 1 .4 Prepositional phrase 11.5 Genitive phrase

126 126 127 131 134 136 137 139 144

83 83 86 86 89 90

Contents vii

12

13

11.6 Verbal phrase 1 1 .7 Subordinator phrase 11.8 Functional analysis of phrase structure

146 149 150

Word structure 12.1 Morpheme, morph and allomorph 12.2 Word structure

151

Functional syntactic analysis

156

151 152

Part IV Complexity and complementation 14

15

16

Complexity (1) 14.1 Recursion 14. 1 . 1 Linear and embedded recursion: the traditional analysis 14. 1 .2 Bound and rankshifted clauses reviewed 14.2 Discontinuity 14.3 Phase

165 165

Complexity (2) 15.1 Apposition 15.2 Thematic focusing 1 5.2. 1 Extrapositional construction 1 5.2.2 Predicated theme / cleft construction 15.2.3 Pseudo-cleft construction 15.2.4 Existential 'there' construction 15.3 Prepositional passives

182

More on complementation 16. 1 Verbal phrase complementation 16.2 Nominal phrase qualifiers and complements 16.3 Adjectival phrase complementation

197

165 169 175 176 182 188 188 189 190

190 194 197 203 210

17 Revised functional syntactic analysis 1 7. 1 Clause structure elements - default modes unmarked 17.2 Clause structure elements - Predicator and Verbal element 17.3 Phrase structure - substructure analysis 1 7.4 Phrase structure - headword analysis 17.5 Phrase structure - phrasal elements 17.6 Postscript

221 221 221

Select bibliography

237

Index

242

224 225 225 235

This page intentionally left blank

Introduction

One of the many remarkable things about language is that we can use it daily without any real awareness of how it is structured. Likewise we can use it in very different circumstances without being at all conscious of the important role played by the particular situation on our choice of language wording. Yet a change, for example, in the social role we are playing or in who we are talking to will typically prompt us to alter, sometimes quite significantly, the actual form of words we use. Indeed, it is by the selection not just of lexical items but also of grammatical structures that we are able to express different meanings. In this way we can begin to point to the link between language wording, meaning expressed and situational context. The theory of systemic grammar was originally formulated by M.A.K. Halliday in the early 1960s. Michael Halliday had been a student of the British linguist J.R. Firth and his early formulation owes much to the influence of Firth's teaching. But as the theory developed, it increasingly began to represent an advance on Firth's own thinking. From its very beginnings, though, systemic grammar has been marked by its recognition that all language, whether it occurs in the form of a book, a letter, a group discussion, a casual conversation, or a person's emotional outpourings, takes place in the context of a social situation, that the situation has a impact on the nature and meaning of the language used and that any account of language must therefore include reference to that context of use. In the first half of the 1960s the theory was known as 'scale and category grammar' (see Halliday 1961). To begin with, attention was focused largely on grammatical structure alone. The theory saw the linguistic system as comprising the level of form, itself made up of lexis and grammar, together with two interlevels, context and phonology. Context provided the link between situation and form, and phonology provided the link between form and sound. The grammar set out to handle the analysis of stretches of language that had actually occurred, so that at this stage the grammar was descriptive rather than generative. Any corpus of written or spoken language material selected for descriptive analysis was known as 'text' and to facilitate grammatical description a framework of categories ('unit', 'structure', 'class' and 'system') and scales ('rank', 'exponence', 'delicacy' and 'depth') was established. Hence the derivation of the name 'scale and category grammar'. During the latter half of the 1960s Halliday's work became influenced by ideas on the functional nature of language. This appreciation of functions of language was not, of course, in itself new. In the 1920s and 1930s Malinowski1 and Buhler2 had discussed the notions of a conative function in which language acts as a form of social control, of an expressive function in which language serves to express a speaker's feelings, and of an ideational/representational function in which language is a means of communicating ideas. These ideas about function now had a profound effect on the nature of systemic grammar, and a multifunctional semantic dimension was not just added to the overall grammatical framework but in fact became central to it (see, for example, Halliday 1967-68, 1969b, 1970b). Resulting from the inclusion of this new dimension, which was presented in terms of three broad metafunctions - the ideational, the interpersonal and the textual - reflecting three types of meaning in language, the grammar no longer sought merely to analyse the syntactic (or, in Halliday's terms, the lexicogrammatical) structure of stretches of actual text. Interest now switched to accounting for the nature of the total system of linguistic meaning available to the native speaker of a language and for the selection of actual options which a person makes when using that language on any particular occasion. These options are thus selected not from the syntax but from the semantics of the grammar, and they represent the choices of meaning which the speaker or writer selects and

2

Introduction

expresses in the context of a given situation. The options are then realized as elements of the language structure, that is to say as the various component parts of the lexical, grammatical and phonological form being spoken or written. In this way the grammar had thus become generative. Simultaneously with this reorientation of goals, the earlier name of the theory, 'scale and category grammar', was replaced by the label 'systemic grammar'. The fuller title is 'systemic functional grammar', which for working purposes is often abbreviated to 'SFG'. Most linguists operating in the systemic tradition nowadays use versions of the modern label to denote all the work - syntactic, semantic and contextual - in the Hallidayan mould since 1961, and that is the line adopted in this book. With the increasing international interest in the functional nature of language and in the linguistic analysis of texts, Halliday is of course today only one of a large and growing number of linguists working within a broadly systemic framework, and writings by systemicists such as Berry, Fawcett, Gregory, Hasan, Martin and Matthiessen, along with many others, have made a substantial contribution to the development of the theory. However, the major contemporary reference work is Halliday's Introduction to Functional Grammar, appearing first in 1985 and with a second edition in 1994. It should be stressed too that one of the theory's hallmarks, as evidenced for example in programmes of International Systemic Functional Congresses, is the extent to which systemicists have applied the grammar to other areas of study, e.g. language generation, literary analysis, discourse analysis, legal analysis, translation, education and language development. In a sense this book goes back to the theory's scale and category roots in that the main aim is to present a descriptive syntactic framework, but now in the light of the overall semantic functional dimension. The contents are divided into four parts. In the first I outline briefly the nature of systemic grammar as a whole, including the linguistic system, context of situation, and language functions. This, though, is intended merely to provide the setting for the main thrust of the book - the lexicogrammar, which concentrates on the analysis of syntactic structure. In the second part, formal units and their classes are outlined, but the principal focus of the descriptive framework is in the third part which is concerned with their role as elements of syntactic structure. Essentially, therefore, the intention is to show how the analysis of syntactic structure can more fully and accurately reflect the meaning structure of the language. To quote Gregory (1987: 99), 'there are a host of purposes, such as stylistics, text description and language pedagogy, for which semantically revealing syntactic analysis has a place'. As an integral aspect of the presentation, approaches by other systemic (and, at times, non-systemic) grammarians are also considered. At each stage the text is illustrated by examples and at the end of the second and third parts ('Class' and 'Structure') the formal and functional frameworks are applied to the analysis of a range of sample sentences. Part IV discusses several areas of structural complexity, and concludes with a chapter which contains some further examples of analysis and which also proposes a number of refinements to the actual presentation of the analysis. The book has two purposes. Firstly, although there are several recent, good introductions to systemic grammar on the market which focus primarily on the semantic functional dimension (e.g. Bloor and Bloor 1995, Eggins 1994, Thompson 1996 and, of course, Halliday 1985a/ 1994a), many lecturers wish to teach the syntactic framework before tackling the semantic dimension. Following this sequence allows them to provide students with the lexicogrammatical terminology to be able to discuss later how semantic features are realized in the wording. At the same time, it is important that the syntactic framework should be both rigorous and well-developed. So, whilst I fully accept the underlying orientation of systemic grammar that it is the choices from within the semantic system networks (that is to say from the total range of options available) that are realized through the lexicogrammatical wording, I would argue that a book on systemic syntax fulfils a role in its own right and, indeed, one for which there is a real need. (Any criticism, therefore, that, in not first developing semantic system networks and integrating their output, the text deals with only half the grammar, completely misses the book's very rationale.) A second and very different reason for the book is that, for the lecturer who wishes to teach a course in syntactic structure on its own, it is important that there should be an alternative to the many contemporary textbook introductions to syntax which are written, to a greater or lesser extent, from a transformational generative perspective. As will be

Introduction 3 seen, by contrast with the tree diagrams which are found so often in linguistics textbooks and which are associated with the representation of structure from a logical dependency perspective, the structural representations used in systemically orientated syntactic analyses present a much flatter appearance. They operate rather on a minimal bracketing basis in that word sequences are grouped in terms of how they function in the next largest unit. This approach to the analysis of structural constituency is one which Halliday (1994a: 22) calls 'functional bracketing' and although it is orientated inherently towards the analysis of functional structure, it can of course be applied also to the analysis of formal constituency. The differences between the two approaches can be illustrated with a contrastive analysis of The cat sat on the mat using non-technical terms: maximal bracketing

A

C

B

D

E

The cat sat on the mat

Here the analysis marks that A rewrites as B and C and that C rewrites as D and E. D and E are thus dependent on node C and only indirectly and jointly, through node C, dependent on A. minimal/functional bracketing

A

B

D

E

The cat sat on the mat

Here the analysis marks that B, D and E each have a direct role in the structure of A. This book is designed principally for students of language and linguistics, and applied linguistics. It should, in addition, be of interest to students of language education and to advanced students of English. The first three parts, on the linguistic framework, class, and structure, are suitable both for intermediate undergraduate and for graduate study, whilst Part IV on complexity and complementation delves into some more advanced topics. Notes 1 B. Malinowski (1923) 'The problem of meaning in primitive languages', Supplement I in C.K.. Ogden and I.A. Richards (eds) The Meaning of Meaning, London: Regan Paul; B. Malinowski (1935) Coral Gardens and Their Magic, vol. 2, London: Allen & Unwin. 2 K. Biihler (1934) Sprachtheorie: die Darstellung der Sprache, Jena: Fischer, and translated by D.S. Goodwin (1990) as Theory of Language: The Representational Function of Language, Amsterdam: John Benjamins.

This page intentionally left blank

Parti

The linguistic framework

This page intentionally left blank

1

The systemic functional framework

1.1 The linguistic system Language represents just one of the ways in which we as humans can behave, that is to say can perform a behavioural act. We can, of course, behave without resorting to the use of language: for example, we can play the piano, mow the lawn or cook the dinner. We frequently behave partly with the assistance of language, as for example when we greet someone by shaking hands at the same time as saying Hello, Jill, pleased to meet you or when we offer to help with the washing up by asking Can I give you a hand? as we pick up a tea towel and start to dry the dishes. But a great deal of our behaviour is performed by the use of language alone, e.g. Can you tell me the way to the Post Office, please?; Would you like a cup of tea?; I'll be there by eight thirty; Do make less noise, will you! Whenever language is involved, we are dealing with 'verbal behaviour'. Whether the language occurs in spoken or written form Halliday has labelled it as text, though, with the increased interest in the nature of spoken language, other writers have employed the term discourse when referring to conversation or indeed any examples of language in the oral medium, preferring to restrict 'text' to occurrences of written language. (Indeed, somewhat confusingly, some linguists have assigned the label 'discourse' the same scope that Halliday has given 'text' - to encompass both spoken and written language.) However, irrespective of whenever and wherever language is used or encountered, systemic grammar holds that it is a social activity which always takes place in a context. The context of situation handles the dimensions of the situation which have a bearing on the language used and is studied under register. Related to the contextual dimensions, language is interpreted as fulfilling a number of different functions: ideational, interpersonal and textual, in that it enables people to express different types of meanings. The ideational function provides for the expression of our experiences of the world as the 'factual' or 'content' element of what we say. The interpersonal function enables us to develop social relationships, to interact with others and to get things done. The textual function allows us to organize what we say or write into a coherent and cohesive piece of text which will both relate to what has already been said and take account of relevant aspects of the context. (See Section 1.3, Language functions and the semantic stratum.) In systemic linguistics the grammar or linguistic system of a language itself is seen as comprising three levels or strata: the semantic stratum, the lexicogrammatical stratum and the phonological (or graphological) stratum. The semantics account for the structure and patterning of the different components of linguistic meaning of a text and, reflecting the different functions which language fulfils, are normally seen as constituting the grammar's generative base. The lexicogrammar accounts through syntax, morphology and lexis for the wording structure and patterning of a text, and the phonology accounts for its sound structure and patterning (or the graphology accounts for the written/printed form structure and patterning). Together the lexicogrammatical and phonological/graphological strata realize the output from the semantic stratum, that is to say they translate the meaning of each of the semantic components into discrete lexicogrammatical and phonological/graphological structures which are then mapped onto one another. A text thus involves the fusion of several different layers of structure, in which the lexicogrammar and phonology/graphology give linguistic form to the semantic output. In turn this linguistic form is given sound (or written) expression through phonetics (or graphetics).

8

The linguistic framework

The basic relationship of the three constituent strata of the linguistic system to context and phonetics is represented in the diagram below.

language system

situational context

register

meaning structure wording structure sound structure

semantics lexicogrammar phonology

spoken expression

phonetics

This diagram may, however, be slightly reoriented to reflect rather better the relationship between semantics and the two realizational strata. The linguistic system:

language system

situational context

register

meaning structure / \ wording — sound structure structure

semantics / \ grammar — phonology

spoken expression

phonetics

The semantic generative orientation of systemic grammar thus stands in sharp contrast to that found in versions of the standard theory of transformational grammar. There it is the syntactic component, with its phrase structure rules, lexicon and transformation rules, which is the generative source of the grammar and the semantic and phonological components fulfil, rather, an interpretative role.

1.2 Context of situation The situational context surrounding acts of verbal behaviour is studied through the parameters of register. Originally based on the thesis that 'language varies with situation' and therefore that 'a certain kind of language is appropriate to a certain use', register was defined as the variety of language used in a particular situational context. In this early interpretation, register was seen as merely providing a framework for describing the background setting for a text. During the 1970s, however, a stronger claim was made that the contextual dimensions of register account for the nature and meaning of the text and can even be said to determine it. Register is, traditionally in systemic grammar, studied in terms of three parameters: field of discourse, tenor of discourse and mode of discourse. In the process of language generation these parameters would be seen as projecting forward to a text, but for someone engaged in textual/ discourse analysis they would be derived rather from the text, with backward reference to it. The field of discourse is concerned with the subject matter of the text, what the text is about, e.g. mountaineering, choral music, gardening, car maintenance, interior decorating, meteorology. In addition to specifying the general nature of the subject matter or topic area, however, field is concerned with what is happening, who is doing what, and how, why, when and where they are doing it. It thus seeks to identify the participants or things involved, the actions and

The systemic functional framework

9

events taking place, and any relevant surrounding circumstances such as the time, location, manner, purpose, etc. For example, the text Whip the cream until it is thick relates to the field of food preparation and concerns the processing (whip) of a food (cream) for a given period of time (until thick). The text Staccato at bar forty-nine, please relates to the field of music and involves the speaker indicating the manner (staccato) of the singing or instrumental playing required at a given point (at bar forty-nine). The tenor of discourse is concerned with the social status and role of the various participants and the relationship between them in the situation. This will be reflected in the degree of formality or familiarity in the wording of the text. It takes account, therefore, of the fact that during the course of a day a person may operate in very different capacities both in respect of their status relative to that of other interlocutors and in respect of their actual role. So, for example, the same person may be a manager giving advice to an employee, someone seeking approval for action from a senior manager, a patient describing his/her symptoms to a doctor, a parent discussing the progress of his/her child with a teacher, a parent discussing a problem with his/her spouse, a parent discussing a school matter with his/her child, a student at a language evening class, a potential customer talking to a car salesman, or one of a group of guests chatting to each other at a party, etc. Together with the question of interpersonal relationships is that of what the language is doing, the purpose of the discourse - its speech function. For example, Could we stop for a minute, please? is a request, Follow the path along the ridge for a mile past the summit cairn is a direction, Would you like a sweet? is an offer, Don't let go of the rope, whatever you do is a warning command, Well done is a congratulation, Hello, ???, pleased to meet you is a greeting. Any text, of course, involves a continuous sequence of speech functions, and even a single sentence may contain more than one, e.g. Could you make a pot of tea, please, and I'll mow the lawn. Traditionally in register studies, the social role or purpose of the discourse was always handled as a subparameter of tenor. Thus Gregory (Gregory and Carroll 1978: 51), for example, distinguishes between personal and functional tenor. A decade later, however, Gregory (1988: 314) specifically renounces his earlier distinction, now believing that too limited a view of communicative function was given by confining it to the interactive relationship. Under this realignment, 'speech function' is accounted for simultaneously alongside all interpersonal systems, including discoursal 'turn', 'social distance', personal 'mediation', etc. The mode of discourse is concerned with the language medium through which the text is expressed. The two primary contrasts are 'spoken', which people might initially associate with 'spoken to be heard' as in ordinary conversation, and 'written' which might be most readily connected with 'written to be read' as with a book, newspaper, letter, e-mail or fax. But between these basic divisions are several intermediary and often interrelated stages. Thus, within spoken language there is a distinction between 'spontaneous' and 'non-spontaneous' language. Spontaneous spoken language is the normal way of characterizing conversation or 'dialogue'. In addition, it can be found in a 'monologue' situation in which participation by more than the person speaking is effectively ruled out either because the speaker is 'hogging the conversation' and not letting anyone else get a word in edgeways, or where someone is giving a running verbal commentary on, say, a football match. Non-spontaneous spoken language requires the participant(s) to learn the material beforehand and then deliver their lines from memory; it would perhaps be most readily associated with the concert monologue when a person recites a poem or tells a narrative tale, such as 'Albert and the lion' by Stanley Holloway. (There is also the possibility of non-spontaneous 'dialogue' in the form of a concert sketch.) Written texts, in addition to being 'written to be read', may be 'written to be read as if spoken', as is the case with dialogue elements in a novel. Or they may have been written specifically 'to be spoken', as for example with the script for a radio or television news bulletin. They may even be 'written to be spoken as if not written', as with the script for a play or television soap opera. Adapting from a useful diagram by Gregory and Carroll (1978: 47), below is a summary of the types of medium which have been mentioned.

10

The linguistic framework

Register mode: type of medium

Together with these discrete types of medium, account must be taken of the possible interplay between spoken and written language. For example, when someone reads aloud a story to a child, it is an instance of 'speaking what is written to be read'. A radio or TV news bulletin as heard by the listener illustrates the 'speaking of what is written to be spoken'. The actual performance of a play or television soap requires the 'speaking of what is written to be spoken as if not written'. The verbatim transcription of a recorded interview involves the 'writing down of spoken spontaneous dialogue'. This interplay can also be handled by distinguishing between 'medium', the mode of the source text as already outlined, and 'channel', the modality in which the hearer/reader actually encounters the language, i.e. the modality of the received output (see also Halliday and Hasan 1989). Channel employs the terms 'phonic' and 'graphic'. By phonic is meant that the language text is encountered orally, i.e. the recipient hears it as speech. By graphic is meant that the recipient encounters the language as reading material. Within channel, account should also be taken of two further aspects which have a bearing on the form of language used. There is interest, on the one hand, in the use of any technological aid, e.g. radio, telephone, audio recorder, television or computer, as a vehicle for the transmission of the phonic/graphic output to the hearer/reader, and, on the other, in the type of textual base, e.g. newspaper, book, brochure, poster or letter. Before going further, however, I would draw attention to the fact that, in addition to the study of context of situation, there has been increasing interest in what is termed the context of culture. This reflects the social and cultural background in which the language is set and the participants' understanding of the prevailing social meanings and cultural values. In systemic linguistics it is studied under genre, which is concerned with the classification of types of social behaviour within a given cultural environment and, using that framework, with the assignment of the function of any given interaction. In determining the purpose of a discourse or text in this way and hence the genre to which it belongs, the linguist is concerned to identify the sociobehavioural activity - through its various component stages - which the participants engage in, as evidenced in the language text. For example, a transactional visit to a bank to withdraw money might include the following stages: A: B: A: B: B:

Next, please. Can I take out £50, please? How would you like it? Four tens and two fives, please. Thank you.

(offer of service) (request) (question) (response) (thanks)

A strong exponent of the study of genre, Yen tola (1988) sets out an overview of the linguistic system which identifies three planes: genre, register and language, and she describes these as representing a 'systemiotic analysis'. The plane of genre handles systems of social behaviour and it is this plane which controls the choices in register, the semiotic plane. The components of register - field, tenor and mode - are then realized by the plane of language. In being thus pragmatically orientated towards establishing the purpose of the discourse/ text, the nature of genre study clearly overlaps with the study of speech function / social role as outlined earlier under tenor. The problem, of course, ceases to exist if the scope of tenor in register is confined to the nature of the interpersonal relations and social distance between the

The systemic functional framework

11

participants, thereby leaving interactional function/purpose to genre. However, not all writers have chosen to operate that limitation - for example Gregory (1988) and, in practice, Halliday and Hasan (1989) have not done so. Added to this, there is the issue of the degree of match between tenor in the register context and the interpersonal component of the semantics. Suffice it to say, the study of a text or discourse does, somewhere and somehow, require the specification of both the communicative purpose (the speech function) and the subject field, interpersonal tenor and role of the language mode. 1.3 Language functions and the semantic stratum As stated earlier, systemic grammar views language as a social activity taking place within a situational context and fulfilling a number of different functions. It was there pointed out that this idea is not, in principle, new, and reference was made to the earlier work of Buhler and of Malinowski. Halliday posits the view that there are essentially three main linguistic functions which adult language fulfils: ideational, interpersonal and textual. These in turn are seen as reflecting the different aspects of linguistic meaning and are accounted for respectively by the three components of the semantic stratum under the same headings.

Ideational function Reflecting the field parameter of register, the ideational function of language is concerned with the communication and interlinking of ideas and may itself be broken down into the experiential and logical functions. The experiential function is the one whereby a speaker expresses the prepositional content elements of his/her utterance, in other words communicates his/her ideas. In operating this function the speaker refers to people, objects and abstractions, actions, events and states, features and qualities, and relationships of location, time, manner, reason, etc. Within the experiential component of the grammar these are grouped and accounted for under the headings of participant / participating entity (relating to people, objects and abstractions - typically nominal), process (actions, events and states - typically verbal), attribute (features and qualities - typically adjectival), and circumstance (relationships of location, time, manner, reason, etc. - typically adverbial). The role relationships between the participants and the process would also be specified. Thus, for example, the experiential content in the sentence John planted a vine in the greenhouse can be said to be: person = John, + action, past = planted, + object = a vine, + location = in the greenhouse, and could be handled in the grammar as follows: (participant) (process) (participant) (circumstance) agent action goal location John planted a vine in the greenhouse. In Your tiny hands are frozen the content is: object, plural = hands, + attribute = tiny, + possession by addressee = your, + state, present = are, + attribute = frozen,

12

The linguistic framework

and would be analysed in the grammar initially as follows: (participant) (process) (attribute) patient state feature Your tiny hands are frozen. This sort of data is handled in the grammar through the systems of transitivity and voice. The logical function relates the prepositional ideas and elements of these ideas to each other on an equal or subordinate basis. It thus encompasses relationships of coordination, subordination, apposition and modification, as illustrated by the following examples. (Actual word insertions in the text whose sole function is to mark these logical relationships I would, however, see as being part not of the logical function but of 'conjunction' in the textual function, often working in concert with the experiential function.) coordination:

Jack and Jill went up the hill. This tune has a good melody line but the harmony is weak. subordination:

When she comes home, she always makes a cup of tea. Out here we burn peat on our fires, which you can't do in London. apposition:

The president, a pilot, is giving a talk on the history of aviation. Their latest rule, that all members must wear a tie, won't last long. modification:

The news of the party's victory came as quite a surprise. The talk that she gave last time was absolutely fascinating. We can then illustrate the interaction of the experiential and logical subcomponents by reference to the above example When she comes home, she always makes a cup of tea: time time agent action

When

she

comes

agent frequency action

goal

location

home,

she

always

makes a cup of tea.

Interpersonal function The interpersonal function mirrors the tenor parameter of register and is evidenced in two main ways: through verbal interaction and exchanges with others and through personal mediation/ modulation of the main idea/content. In terms of social interaction, language serves to establish and maintain social relations, as is found in greetings and various forms of phatic communion. Thus, in everyday conversation Good morning and Hello. How are you? and Nice to meet you are typical examples of greetings, serving to open or further social contact. Normally such expressions are used to make such contact on a very superficial level, and reference to the weather or to a person's health is purely a way of achieving the initial acquaintance. Thus, for example, as a social greeting How are you? is not a serious enquiry into the state of health of the person being addressed. Irrespective, therefore, of that person's actual condition it would be usual, especially where the participants

The systemic functional framework

13

are not well acquainted, to reply not with a comprehensive medical account but rather something along the lines Very well, thanks. And you? or Not so bad, thank you. The exchange of utterances in the social function thus follows very much a formulaic pattern which can be anticipated, and the language used merely serves to pave the way for more substantial discourse. Indeed, someone who in reply to a comparative stranger did break this pattern and embark on a full account of his or her ailments would quickly be marked out as a bore. It is, however, quite possible for phrases such as Nice day or Lovely weather, isn't it? to provide the opening for a fairly lengthy conversation about the weather. But in such circumstances the conversation could well remain at a superficial level because the topic is still being used as one which can be 'discussed' non-controversially and which will thereby help to remove the barriers to further social interaction. With respect to instrumental interaction, language may be used to seek to influence the behaviour of others in order to get things done. It accounts for the fact that through the use of their chosen 'illocutionary force' people issue commands, make requests and put forward suggestions to direct people's actions/behaviour; they offer goods to the addressee, offer to perform a service for him or her; and they express their will/wishes. Some of the manifestations of this function are illustrated below. command:

Keep out. Turn left at the lights.

request:

Will you pass me the teapot, please? Could you lend me a pound? May I borrow the map, please?

suggestion:

How about making pancakes for tea? What about trying a different brand?

offer:

Would you like a cup of tea? Can I give you a hand?

will/wish:

Let this be a lesson to you! May you never do this again! Long live the king!

With informational interaction people give or seek (content/factual) information. They make statements to impart information and ask questions to seek information; they utter exclamations (incorporating some content element) by way of reactive comment. statement:

The Chancellor has abolished income tax. Today only one train was late.

question:

How much will it cost? What's the time? Was it expensive? Will you go by car or take the train?

exclamation:

What a good idea! How stupid!

Closely linked to the idea of reactive comment through exclamation, through expressive interaction people can, again through exclamations, give vent to their emotions, but here without a content element, e.g. Heavens above! Good gracious!

14

The linguistic framework

Through personal mediation/modulation the speaker can also express his/her own personal attitudes towards the ideational content of what is being said (and his/her personal assessment of it). These reactions and comments by the speaker thus serve to modulate in a subjective way the mainstream idea or proposition in the utterance, as seen in the following examples: Frank has probably got it. Maybe the letter didn 't reach her. She may never have received it. She must have received it. You 've surely finished with it by now. Unfortunately the parcel never arrived. Astonishingly Max came back after only two days. Thankfully she didn't stay long. Wisely they didn't wait around. Also the speaker can set out the terms of reference in which he/she is speaking: To be frank it's not worth the trouble. Honestly it was great! Essentially all you need is extra RAM. From an economic point of view the situation is improving. En vironmentally it could be a disaster. Travel wise we 've got three possibilities.

Textual function The textual (or discoursal) function is described by Halliday as the one whereby language serves as a means to create texts as opposed to merely isolated and disconnected sentences. It is the function which organizes the language in a textual corpus in such a way as to give it narrative coherence (in which the ideas are presented in an acceptably logical sequence) and message cohesion (in which the wording of a sentence in a discourse takes account of and is linked to that of previous sentences), to arrange it as units of information, and to avoid unwanted redundancy. It can be said to be concerned with shaping the nature of a text in its spoken or written mode, in other words with fashioning the texture of a passage. The scope of the textual function thus extends beyond individual sentence boundaries. Through the textual function the speaker is, firstly, able to give a thematic structure to the elements of the clause content, thereby highlighting one or other element in first position and giving it thematic prominence. Compare, for example, the following: They bought the computer yesterday. The computer they bought yesterday. Yesterday they bought the computer.

(Agent they as theme) (Goal the computer as theme) (Time yesterday as theme)

Secondly, the speaker organizes the clause content in terms of information units. Each of these contains an element which is being presented as new information in respect of previous discourse and may also contain an element of given information, which is being presented as being recoverable from previous discourse. Then through the use of phonological stress, the speaker centres attention on one (or more) part(s) of each information unit as the information focus, e.g.

The systemic functional framework

15

They bought the computer yesterday. (Normal, unmarked interpretation. Job done.) They bought the computer yesterday. (i.e. but not the printer or the scanner.) They bought the computer yesterday. (i.e. they didn't hire or steal it.) They bought the computer yesterday. (i.e. it wasn't me or you who bought it.) Thirdly, he/she is able to use pro forma expressions to refer to one or more elements previously mentioned in the discourse without having to repeat their information content in full. In this way language avoids sounding like a robotic product. See for example the following: (i)

Have you got your map of the Lake District? Yes, it's in my coat pocket.

(ii) Last year we went to Salzburg in Austria and travelled by car. Oh, so did I. (iii) Ruth is going to the new museum by the art gallery. Can I come? I've never been there. Compare these responses above, which make use of pro forma expressions, with the ones below, which merely repeat the reference from the previous utterance in full, thereby creating a very unnatural effect: (i)

Have you got your map of the Lake District? Yes, my map of the Lake District is in my coat pocket.

(ii) Last year we went to Salzburg in Austria and travelled by car. Oh, last year I went to Salzburg in Austria and travelled by car. (iii) Ruth is going to the new museum by the art gallery. Can I come to the new museum by the art gallery? I've never been to the new museum by the art gallery. Fourthly, through ellipsis a person can omit entirely repeated mention of those elements which he/she considers to be recoverable from an earlier part of the discourse, e.g. I'm not available. Are you ? Will you be in the cellar? No, in the loft. I can't understand his reluctance. Neither can I Have you got the tickets? No, not yet.

.

As with the use of pro forma expressions for reference, through this streamlining in the degree of mention of elements of message content, ellipsis enables the speaker to avoid redundancy in utterances and hence again a possible robotic quality. Lastly, within the textual function, through the resources of conjunction (a broader concept than 'conjunctions') the speaker can insert words and phrases to mark different types of cohesive relationships between clauses and sentences, e.g.

16

The linguistic framework

She is, however, a good judge of character. By the way John won't be there tonight. On the other hand you could hire a minibus. Consequently he lost his place. By thus linking the ideas being expressed and minimizing the amount of redundant material, speakers are able to create unified texts. Indeed, without such resources, many texts would be intolerable to read.

The semantic stratum The meaning structure associated with each of the different language functions above is technically accounted for by the semantic components of the grammar. The outline here broadly represents Halliday's schema, in which three components are recognized: the ideational component with its experiential and logical subcomponents, the interpersonal component with interactional and personal subcomponents, and the textual component comprising thematic, informational and cohesive subcomponents. The ideational component reflects the field of discourse from the situational context and is concerned with content meaning input and patterning. It deals with systems such as transitivity and voice. The interpersonal component mirrors the tenor of discourse and handles all aspects of the speaker's interpersonal meaning input and patterning. It includes the interactional meaning of the language used, i.e. the meaning of what the language is doing, in terms of its illocutionary force. (This is distinct from the contextual/speech function of the language - the purpose to which the speaker is putting it.) The interpersonal component also deals with the speaker's personal attitudes, comments and terms of reference, together with interpersonal aspects such as attention-getting frames, modes of personal address, greetings and politeness expressions. This component thus embraces those elements which have either a 'modulatory' impact on the ideational content or a mediating role in respect of other participants in the situation. The textual component deals with the message organization and cohesion in the text, e.g. conjunction, thematic and information structure, reference and ellipsis. An important feature regarding the organization of this semantic stratum is that although each of the components is discrete and distinct in its own right, they all contribute to the overall meaning and structure of the text. They are, further, seen as doing so simultaneously, without any sense of priority being accorded to one or other of the components. Indeed, in most sentences more than one meaning component will be found to be involved. Let us take a look, for example, at meaning elements in the sentence John must have worked very hard at his project because he got a distinction. Here a crude, first stage analysis may be given as follows: ideational:

John ... have worked very hard at John's project; because; John got a distinction.

interpersonal: giving information: must. textual:

selection of John as first element of sentence; referential (anaphoric) use of his in main clause; referential (anaphoric) use of he in subordinate clause.

In John must have worked very hard at his project, the factual content is specified in the form of who (John) performed the process (work), in respect of what (project), and in what manner (very hard). In because he got a distinction, the reason for this is given. In interactional terms, the

The systemic functional framework 17 sentence serves to give information as opposed to seeking it or even to directing anyone's behaviour. In personal terms, the speaker concludes (must) that the information is so. Textually, the mention of John as the first element in the sentence highlights him thematically as the entity being talked about; the use of his and he in place of John's full name helps give cohesion to the passage. Also the word because binds the two elements of the sentence in a logical and causal relationship, thereby giving the passage coherence. The standard Hallidayan arrangement of the components of the semantic stratum and their relationship to the components of the register context can be represented as follows: context

field

tenor

mode

semantics

ideational

interpersonal

textual

It should be said, however, that not all systemic linguists are agreed upon this division of the semantics into three components. Perhaps the most divergent is Fawcett (1980), who delineates eight main components: experiential, logical, negativity, interactional, affective, modality, thematic, informational, and three minor ones: inferential, metalingual and discourse organizational. Earlier Fawcett had handled 'negativity' under the 'logical' component, and 'affective' (the speaker's evaluation) and 'modality' under the 'expressive' component (which was roughly equivalent to Halliday's personal subcomponent). The 'metalingual' is found 'when the performer stands back from his message' and interpolates an appraisal pause check on it, e.g. so to speak, as it were. Within 'discourse organizational' Fawcett includes conjunctive cohesion, 'attention-getting frames', e.g. Right then, Excuse me, and the use of the 'vocative', e.g. Fred, could you — Ventola's overall linguistic framework (1988) is built around three planes - genre, register and language. Her plane of language is organized on three strata: discourse, lexicogrammar and phonology. The level of discourse handles reference, lexical cohesion, conversational structure and conjunction. For her, the lexicogrammar encodes the structure of a clause as representation, as interaction, and as message, and is the level where the systems of transitivity, mood and theme operate. (The phonology, of course, encodes the phonological patterning.) Interestingly, though Ventola's outline does not actually include a level of semantics as such, it still operates on the basis that the lexicogrammatical structures of the clause 'realize, in the Hallidayan framework, the ideational, interpersonal and textual metafunctional choices of semantics' (1988: 59). 1.4 The systemic orientation The name 'systemic grammar' is derived from the fact that a language is seen as being a huge, integrated series of systems networks of meaning potential. This represents the generative potential of the language, and it is the task of the grammar to specify this total grid of options available to a speaker. The situational context provides the semiotic parameters and constraints, including the interactional purpose of what the speaker 'is doing' behaviourally in using the language. The semantics accounts for what the language 'can mean' and, within that, what the speaker 'is meaning'. The lexicogrammar then specifies the word form which the meaning can take, what the speaker 'can say' and, in any one instance, what he/she 'is saying'. What he/she does actually say on any occasion therefore reflects the way that he/she operates or actualizes the language's potential. In a broad sense the grammar seeks, through the semantic networks, to display the meaning potential which the speaker can utilize and, through the lexicogrammar, to indicate the wording which a given meaning may take. However, unlike the notions of competence and performance in transformational grammar, systemic grammar draws no distinction of principle between the meaning potential and its actualization or use in a given

18

The linguistic framework

context. Meaning potential is thus not seen as some sort of abstract deep structure requiring transformations to form the concrete surface structures that we in fact employ. Actualization of the potential is rather a matter of the speaker making a selection from the range of options which the language has to offer; it is a grammar of use. Although the orientation behind this book is descriptive and analytical rather than generative, I shall now briefly outline the basic principles behind network systems. A system is composed of one or more sets of features or 'terms' in contrast with one another and organized in a pattern of relationships ranging from the most general to the most detailed. The list of terms within a system is finite and the options are mutually exclusive. The meaning and scope of any one term is so interlinked with that of each of the other terms that if a new term is added to a range of options, the meaning and scope of at least one of the existing terms is affected. As a crude analogy, a system accounting for a particular area of grammar may be likened to the number of pieces into which a cake is cut; the size of the cake remains constant, but the size of each slice depends on the number of portions into which the cake is divided. A system can, however, also be developed through the establishment of subsystems, but these further options would not change the form of the initial system. Each system has as its point of origin a specific rank of grammatical unit, e.g. clause or word (see Section 2.2, Units and the rank scale) and may operate with hierarchical or simultaneous entry conditions. hierarchical:

The symbol [ denotes disjunctive choice (either/or). simultaneous:

The symbol (indicates simultaneous choice (both/and).

The systemic functional framework

19

The entry conditions to a point in a system may be as follows: (a) simple - requiring, for example, that only feature x need apply before a further choice between a or b is made.

(b) compound/conjunctive - requiring, for example, both features x and y to apply before the further choice between a or b can be activated.

(c) disjunctive - requiring, for example, either feature x or feature y to apply before the further choice between a or b can be activated.

A systemic specification thus appears as a set of features selected from the range of options drawn from the different networks which together constitute the grammar. Ideally a system should be drawn up to account for similarities and differences in such a way that the features develop from the most general to the most particular as the system progresses further from the point of origin. To illustrate the nature and operation of a systems network, I set out below a simplified partial system of illocutionary force from the interpersonal meaning component.

20

The linguistic framework

Partial system of illocutionary force • giver (1) (statement) information

seeker— (question)

polar (2) (seeking a 'yesY'no' reply) 'wh'-type (3) (seeking new information) disjunctive (4) (seeking choice from information options given)

simple (5) jussive (command)

personalized (6) pressing (7)

directive request (8) self-inclusive (9) (exhortation) suggestion (10) Below are some sentence examples illustrating the output generated by the various combinations of terms: 1. 2. 3. 4. 5. 6. 1. 8. 9. 10.

JiU'll wash it. WillJill wash it? What willJill wash? Have you washed this or this? Wash it. You/ Jill wash it. Do wash it. Will'/can you wash it? Let's wash it. What about washing it?

However, as stated, this example is included purely as an illustration. The intention of this book is not to examine how syntactic structures can be generated from semantic functional networks. The aim is rather to outline and develop the syntactic analytical framework in the light of that semantic functional dimension and thus to show how the analysis of syntactic structure can more fully and precisely mirror the meaning composition.

2

The grammatical framework

2.1 The nature of the lexicogrammar Lexicogrammar in systemic linguistics is the stratum which handles the wording of a text. It is the level at which the various patterns of semantic structure are realized in word form and is concerned with the classes of grammatical (and in principle also lexical) unit and the relationships which may be established between them. This book concentrates on the lexicogrammatical component and in particular on syntax, which in linguistics is traditionally defined as the grammar above the word or the grammar of sentence structure. Relatively little attention, therefore, will be paid to morphology. Notwithstanding the generative nature of the link between semantics and syntax in a fully-fledged systemic grammar, the focus here is on the development of a framework for the analysis of the syntactic structure of a text rather than on semantic networks and the realizational rules which map them into the lexicogrammatical wording. Language is an open-ended organism. Only by recognizing this can we account for the fact that it does not stand still - indeed has never stood still - but is constantly evolving. Any descriptive grammar of a language is therefore merely a snapshot in history. There can thus be no definitive grammar of a language which can be held up as a point of reference for all time. Indeed for parents and grandparents it can sometimes be a point of contention that 'young people today don't speak the way I was taught when I was at school'. Although typically made in a judgemental manner, from a factual point of view the observation is perfectly correct; the way people speak and write now is different from that of 25,50 or 100 years ago. To a greater or lesser extent this always has been the case and with each successive generation it will continue to be so. Consider, for example, people's differing reactions to It is I versus It is me or To whom were you talking? versus Who were you talking to?. It is to grammars of a language from different ages that we look to record the accepted norms and acceptable forms of the day. (In this context we wait to see what history makes of the more recent instance between you and I versus between you and me.) The study of grammar exists at all because a language does not consist of a fixed number, say 100,000, of possible sentences. On the contrary, owing to the creative potential of language, we can produce utterances which we have never heard before and, likewise, can understand sentences which we may not have come across previously, e.g. / will check the speed of these spiders with John's digital watch. As an integral aspect of this creative property of language, we form sentences not by stringing words together in a random, haphazard way but by following certain patterns or rules of structural composition. Thus, for example, we can say / want a banana, perhaps even 1A banana want /, but not */ banana want a, still less *A want banana I. (An initial question mark (?) is used to mark a doubtful structure or occurrence, and an asterisk (*) denotes a non-grammatical one.) It can also be shown that in the structure of sentences some words have a closer relationship to each other than to other words in the same sentence. Put another way, we find that words can be grouped together to form larger units. In fact, several possible groupings of different sizes of unit can be identified between 'word' and 'sentence'. For instance, in / want a banana we point to a close relationship between a and banana but not between / and a or between want and a; there is, however, a relationship between / and want and between want and a banana. Grammar then provides us, the language user and the language learner, with a basis for understanding how a language is structured, what the possibilities of patterning are and indeed

22

The linguistic framework

what the constraints are. It is not just a matter of a person declaring that you can't say *A want banana I because the grammar says you can't. The position is rather that *A want banana I is the type of structure which speakers of the language don't produce for very good reasons of language patterning and relationships. It is the grammar which accounts for these regularities of language structure and which thus explains why speakers avoid such malformations. In this way it can be seen that grammar is 'language led' rather than the other way round. It reflects the way in which language is used / can be used. If, to take the reverse position, a language had been 'grammar led', then grammar would have determined the way we use the language and, given authoritarian enforcement, its nature might not have changed over the centuries and millennia in the extensive way that it has. (This is not to say, of course, that we might not from time to time have recourse to a reference grammar specifically for guidance on the form of a particular grammatical norm.) Differences of wording in a text reflect differences of meaning. Accounting for differences of meaning is, as has been said, the job of the semantics. It is the semantics, therefore, which would handle, for example, the meaning contrasts between the following, related sentences: I'll give John the map; I'II give the map to John; Shall I give John the map?; Give John the map!; John has been given the map; The map has been given to John; John is fighting Bill; John is fighting with Bill; John has turned off the light; John has turned off the road. It is then the task of the lexicogrammar to account for the differences (and the similarities) in the actual wording of these sentences, that is to say differences and similarities of syntactic structure. Syntactic structure can be approached in two ways: formal and functional. Formal syntax deals with how words can combine to create larger units of form and eventually sentences. One can perhaps visualize this as an orientation towards unit building, with a progression upwards from the word to the sentence. This bottom-upwards perspective reflects the question 'What increasingly larger, formal units can we build up with words?'. Functional syntax, on the other hand, handles the way in which sentences are structured in terms of smaller functional elements and eventually words. (A functional element marks the syntactic role that a unit is playing in structure. For example, although a clause is determined as the next largest formal unit to a sentence, it may function as an element of structure not only of a sentence but also of a unit smaller than a sentence.) One can see this as an orientation towards structural decomposition, in which there is a progression downwards from the sentence to the word. This top-down view of the situation reflects the idea that, in terms of what an utterance is doing, a functional syntactic analysis is interested firstly in the question 'What (ever smaller) functional elements can we break sentences down into?' and only secondly in its formal composition. Witness, for example, the case of the native speaker of a language who goes into a self-service cafe to order him/herself a cup of coffee and says merely Coffee, please. This doesn't lead to a breakdown of communication because the speaker didn't utter the more fully formed I'd like a cup of coffee, please or even Could you please sell me a cup of coffee. Witness, too, the efforts of the visitor abroad who knows only a smattering of the foreign language but who in a shop doggedly displays his/her, possibly monosyllabic, linguistic prowess to purchase the goods in question. In such a context it is thus not completeness or even accuracy of formal structure that is the primary consideration. That does not mean, though, that the formal composition is to be disregarded. Indeed, an educated native speaker of a language requires not just to be able to get things done by the use of language (the get-by approach) but should be able to do so by having recourse to linguistic structures which are well-formed and appropriate to the context. Formal and functional syntax are thus seen as complementary: they have different basic concerns but they are interdependent. The lexicogrammar, therefore, will incorporate both the formal and functional features of syntactic structure. As mentioned in the Introduction, in its early stage the lexicogrammatical framework for systemic grammar was developed against a background of four categories: unit, structure, class

The grammatical framework 23 and system, and three (or four) scales: rank, exponence, delicacy (and depth). In subsequent chapters, 'class' and 'structure' will be discussed in detail, but we offer here a brief overview of all the categories and scales, in other words, of the 'scale and category framework'. Essentially the categories are theoretical constructs, whereas the scales serve to relate the categories both to each other and to the textual data. The category of unit accounts for stretches of language of varying lengths and composition, such as 'sentence' or 'word', which carry grammatical patterns or which operate in grammatical patterns. The different units are set out on a hierarchical basis and are related to each other through the scale of rank. Thus, for example, the sentence belongs to a (much) higher rank than the word. In accordance with the category of class, members of a unit are assigned to a particular class according to similarities and differences of their grammatical behaviour. In these terms, 'noun' and 'verb', for example, can be established as two classes of word. The category of structure accounts for the composition of a unit in terms of the functional elements it consists of as well as for the relationship between these elements. Thus, for instance, in a sentence such as Will you ring or will you wait till I come one might consider (a) Will you ring, (b) or will you wait, and (c) till I come as elements of sentence structure and then examine the different relationships between them. Through the scale of exponence the categories are related to each other and to the textual data. Proceeding down this scale also involves regularly changing rank. So, for example, a sentence unit would be said to have a structure in terms of one or more functional elements, each of which is expounded (or - using the more contemporary term - realized) by a unit of the rank next below. Eventually, when for each element the scale of exponence has completed its passage down the rank scale, the process then goes on to specify the lexical item by which each ultimate unit is expounded/realized, that is to say, it moves out to lexis. The category of system in lexicogrammar operates in the manner already illustrated in relation to semantics and is able to account for the choices (classes and subclasses) available within each unit. Instead of just listing the range of options, through system the analyst can set out the options in terms of the relationships of similarity and difference between them. The scale of delicacy is concerned with the degree of detail which is incorporated into the analysis. The scale of depth, on the other hand, if included at all, considers the degree of complexity of a stretch of language in terms of the number of stages by which it is removed from what is termed the point of origin or the source / top node. For example, in the sentence That nest in the corner of the shed by the lilac bush belongs to a robin the phrase by the lilac bush is part of a fairly complex structure of which nest is the head element, the point of origin, from which it is several stages removed, thus: That nest in the corner of the shed by the lilac bush I would make the point, however, that nowadays this framework of scales and categories is not invoked in the same strict manner as formerly. Nevertheless, the basic principles behind it still apply and underpin the operation of the lexicogrammar, even if not overtly. 2.2 Units and the rank scale Syntactic structure in systemic grammar is traditionally based around five formal units: sentence, clause, group, word and morpheme. (Henceforth, as a matter of terminology we shall, like Hudson (1971), use the term 'phrase' in place of'group' to refer to the intermediary unit between clause and word.) These five grammatical units can be illustrated as follows. Sentence \\\After John has finished his exams, he is planning a trip abroad.\\\

24

The linguistic framework

Clause \\After John has finished his exams, \\ he is planning a trip abroad.\\ Phrase \After | John \ has finished \ his exams, \ he \ is planning \ a trip abroad.\ Word : After:John:has:finished:his:exams,:he: is:planning:a'.trip:abroad.: Morpheme -After-John-has-finish-ed-his-exam-s,-he-is-plann-ing-a-trip-abroad.The use of vertical lines for sentence, clause and phrase in systemic grammar is thus: three lines are used to separate sentences, two for clauses and one for phrases. (In syntactic analysis, word boundaries are merely indicated by the normal space divider and morphemes are not involved.) However, the normal notational practice for indicating boundary markers operates on a hierarchical basis such that the insertion of a boundary marker for a larger unit obviates the need to also mark a smaller unit boundary which occurs at the same point. Thus, the presence of a sentence boundary label means there is no need to mark a clause, phrase or word boundary there, and the use of a clause boundary marker similarly dispenses with the need to mark the phrase or word boundary. Consequently, in a proper syntactic analysis the sentence, clause and phrase boundaries would be recorded not as: UN! \After | John \ hasfinished\ his exams, \\\\he\ is planning \ a trip abroad.\ \\\\\ but rather as follows: I I I After \ John \ has finished \ his exams, \\ he \ is planning \ a trip abroad.\\\ A unit is defined as a stretch of language which itself carries grammatical patterns or which operates in grammatical patterns. The sentence is seen as the largest of the four units carrying grammatical patterns: sentence, clause, phrase and word, and the morpheme is the smallest of the four units operating in grammatical patterns: clause, phrase, word and morpheme. Thereafter each of the units is described in terms of its relationship with the other units. All five units are arranged hierarchically, from the largest down to the smallest, on a scale of rank. Each unit except the largest has been defined by its function in the structure of the unit next above, and conversely each unit except the smallest has been described as being composed of one or more units of the rank below. Thus a word, e.g. exams, finished, after, consists of one or more morphemes, viz. exam\s, finished, after; a phrase consists of one or more words, e.g. John, has finished; a clause consists of one or more phrases, e.g. After\John\has finished\his exams; and similarly a sentence is described as consisting typically of one or more clauses. In terms of the way systemic grammar has traditionally been formulated, all utterances require to be described at each rank. Thus, the sentence After John has finished his exams, he is planning a trip abroad would be analysed as consisting of two clauses: After John has finished his exams, and he is planning a trip abroad, of which the first clause contains four phrases: (i) After, (ii) John, (iii) has finished and (iv) his exams, comprising six words and eight morphemes (includingj^wwA and ed, and exam and s), and the second clause contains three phrases: (i) he, (ii) is planning and (iii) a trip abroad, six words and seven morphemes (including plann and ing). But it also means that a reply No! is described as a sentence consisting of one clause containing one phrase composed of one word which is itself a single morpheme.

The grammatical framework 25 The scale of rank sentence clause phrase word morpheme However, the structure of a unit at one rank is not always composed of units from the rank below. For example, whilst the phrase the girl does consist of two units of word rank, the phrase the girl in the corner consists of the same two units of word rank plus the sequence in the corner, which is itself a phrase unit. In other words we have a phrase operating within a phrase. The structure of the phrase the girl who beat the record involves again not just the two units of word rank but also the sequence who beat the record, which is in fact a clause. In this instance, therefore, we have a clause operating within the structure of a phrase. This phenomenon whereby the structure of a unit of one rank is realized not merely by units of the rank below but (also) by units of its own rank or even units from the rank above is traditionally known in systemic syntax as rankshift. This is a concept which, whilst it will crop up incidentally from time to time in the exposition, will be discussed in greater detail in Chapter 14, Section 14.1, Recursion. 2.3 Sentence, clause, phrase, word and morpheme Orthographically the sentence begins with a capital letter and is terminated by a full stop. With the increased interest, also, in the nature and structure of text as distinct from grammatical form alone, the sentence has come to be regarded as an element of textual structure and as such may be seen as a constituent of the paragraph. Strictly speaking then, the sentence is a textual unit which it has been convenient to adopt as the largest grammatical unit for the purposes of syntactic analysis. In meaning terms, the typical role of the sentence is to express one or more ideas or 'propositions' from the ideational component, each proposition being realized by a clause. Indeed, it is very much a matter of the individual writer's style how many propositions, with the help of commas, semicolons and colons, are incorporated into a single sentence. Compare, for example, the extremes in the following two texts: (a) This is the maiden all forlorn that milked the cow with the crumpled horn that tossed the dog that worried the cat that killed the rat that ate the malt that lay in the house that Jack built. (b) This is the maiden. She is all forlorn. It was she who milked the cow with the crumpled horn. That was the cow that tossed the dog. It was that dog that worried the cat. The cat had killed the rat. The rat had in fact eaten the malt. The malt had lain in the house. The house had been built by Jack. In the first text, the traditional rhyme is presented as a very complex sentence (see Chapter 14, Section 14.1, Recursion) comprising an ongoing sequence of propositions embedded inside one another. Potentially, other than that the writer/speaker might run out of ideas or, perhaps, breath, there is no reason from the point of view of the sentence structure why the rhyme sequence should not continue on - and on - still within the framework of a single sentence. In the second text, the sequence has been broken up and each proposition has been presented within its own sentence. From totally opposite perspectives, therefore, both passages have an artificiality about their format.

26

The linguistic framework

I have already mentioned that, in terms of the rank scale hierarchy, a sentence consists of one or more clauses. These are interrelated with one another on a basis which is grammatically coordinate or subordinate and, in consequence of this, some grammarians prefer to think of the largest grammatical unit as being a grouping or complex of clauses, a 'clause complex', rather than a sentence. In fact, the classification in traditional grammar of non-simple sentence types under the three headings 'compound' sentence (e.g. ||| I've finished that book \\ and written my essay.\\\), 'complex' sentence (e.g. \\\When you've finished your tea \\ can you give me a hand?\\\) and 'compound-complex' sentence (e.g. \\\Ifyou want a lift \\just give me a ring \\ and I'll come round.\\\) illustrates a crude, if practical, labelling of different modes of clause complex. A clause in meaning terms typically expresses a single proposition. Grammatically, it consists of one or more phrases, e.g. / 've finished that book. As we shall see, clauses do not all share the same grammatical status, and relationships of subordination and superordination can be established within groupings/complexes of clause units. Consider, for example, the relationship between can you give me a hand and when you've finished your tea in the sentence Can you give me a hand when you've finished your tea? Equally, relationships of fixed and free sequencing can be found between clauses which do share the same grammatical status. For example, in the sentence \\\Just give me a ring \\ and I'll come round \\ or Jill will call\\\, provided we leave the linking words and and or where they are, the second and third clauses could be switched around, but the first clause must stay in its initial position. A phrase in meaning terms expresses one of the elements of a proposition. Grammatically, it is the grouping of one or more words which together fulfil the role that in other circumstances might be expressed by a single word. So, for example, the fastest winner of the 100 metres heats could be replaced by he or she, might have been training could be substituted by trained. Indeed, as we have already mentioned and will discuss further in Chapter 4, Phrase class, the traditional systemic term is 'group' rather than phrase, though Halliday actually distinguishes between 'group' and 'phrase'. The word is the basic unit of syntax. Orthographically, words are typically bounded by a blank space either side of them, as for example in a bunch of flowers. Compound words like birthday and sunshine are still individual units, but the position is complicated by the fact that the same expression may be written as a single, compound word with or without a hyphen or as two separate words, e.g. airlock, air-lock, or air lock. The normal practice of writing words with a blank space before and after them is violated in cases such as Ruth's out, where Ruth belongs to one word class and 's, which here is a contracted form of is, belongs to another. The 's is a clitic, a word form which cannot stand on its own and which, in this instance, is fused or leeched onto the preceding word. Indeed, as a result of their belonging to separate word classes, if one substitutes the first component Ruth by the pronoun They, then the second component 's also has to change to 're. See also They've^ disappeared (contracted have), He'J[play (contracted will), Shejd object (contracted would) or / didn^ know (contracted not). So in these various instances the sequences Ruth's, They've, He'll, She'd, and didn't are handled as two words. The word, then, is a unit which can be assigned to a recognized word class and which is not a (hyphenated or unhyphenated) component of a compound unit. In accordance with this, the analysis of They've and didn't, etc. above as two words is shared by black bird, air lock and computer game, whereas units such as blackbird, airlock, air-lock and home-made are handled as one word. It also means that forms such as girls, works, worked, working remain analysed as single word units because the elements s, ed and ing serve to mark variant grammatical forms of a given word unit and cannot themselves be assigned to word classes. At the bottom end of the rank scale, the morpheme is the smallest unit of grammatical form and meaning - though in traditional grammar a distinction is often made between morph and morpheme. (See Chapter 12, Section 12.1, Morpheme, morph and allomorph). The morpheme is involved in word formation in a number of ways: (a) lexical compounding, e.g. blackbird, sunshine; (b) lexical derivation of one word (or rather lexical item) from another, often involving a change of word class, e.g. act, action, active, inactive, proactive, activate, actor, actress; (c) grammatical inflection altering the form of a word to fit the grammatical context but not thereby changing the word class, e.g. car - cars (plural), mend- mended (past tense).

The grammatical framework 27 With regard to grammatical meaning as illustrated under (c) we can say that a change of morpheme inflection within a noun, for example, leads to a change in the syntactic relations into which the word enters. But unless the morpheme is a free form and therefore able to function as a word, it does not by itself enter into syntactic relations. A morpheme can thus be described as having syntactic significance, but it is not in its own right a syntactic unit. The point was also mentioned above that in grammatical terms the sentence and the morpheme, unlike the clause, phrase and word, have only a one-way relationship: the sentence downward to lower units and the morpheme upward to higher units. With respect to syntax, therefore, the central grammatical units may be delimited as clause, phrase and word, and the primary interest in this book is with these three units.

2.4 Unit complex and complex unit In addition to the basic grammatical units we introduce the concepts of a 'unit complex' and a 'complex unit'. Berry (1975:95) sees these terms as merely alternative labels but we would argue that there is a significant difference between them. When considering the sentence, we made reference to the term 'clause complex' and that can be seen as a model for a unit complex more generally. By contrast with a simple unit, which consists of a single exponent of a unit from a given rank, a unit complex may be explained as a coherent grouping or configuration of two or more units (a complex of units) from the same rank, which together may also constitute a unit of the rank next above. For example, other words may be grouped round a given headword, as in a brilliant student, to form a word complex which is a phrase. (This is different from Huddleston's early formulation (1965a), in which with the example of three nominal phrases in apposition he sees a unit complex as a grouping of units from one rank which form a unit intermediate between that rank and the one above.) If another unit of the same status is involved, then the relationship between them is one of equality, in other words a relationship involving units of equal grammatical status. In systemic grammar this has normally been referred to as one of the types of paratactic relationship. The sentence Jack fell down and Jill came tumbling afterexemplifies a unit consisting of two main clauses coordinated in one such relationship. In Frank Jones, the butcher, is the best candidate the type of paratactic relationship between the units is one of apposition, in which two phrases, Frank Jones and the butcher, are juxtaposed and thus, here, operate as alternative expressions referring to the same person. Where a subordinate unit is involved, the relationship between the main and subordinate units is known as a hypotactic relationship. A hypotactic relationship thus involves units of unequal grammatical status, as illustrated by the subordinate and main clauses in the sentence When Jack fell down, he broke his crown. Traditionally in systemic grammar paratactic and hypotactic structures have been grouped together as subtypes of univariate structure, that is to say a structure involving the repeated utterance of the same type of unit, where the relationship between the units is a logical one of equality or dependence. Where, on the other hand, the relationship is one between different types of element, as for example between the units in Jack Sprat could eat nofat ('subject', verbal element, and 'object'), it represents the type of structure known as multivariate. (For further discussion of types of structure, see e.g. Halliday 1965). In these terms, therefore, a clause complex comprises more than one clause unit; it is a complex of clauses and typically operates as a sentence. A phrase complex involves more than one phrase unit (same or different classes) forming a coherent grouping which in some circumstances can also serve as a clause. A word complex is more than one word unit and can form a phrase. A morpheme complex - more than one morpheme unit - normally forms a word. Below are some examples of unit complexes. Clause complex: Jack fell down and broke his crown and Jill came tumbling after. (paratactic, coordination)

28

The linguistic framework What I really want, what I must have, is a new car.

(paratactic, apposition)

After Keith was awarded his degree, his confidence increased enormously. (hypotactic) Phrase complex: Goldilocks and the three bears (paratactic, coordination) Frank Jones, the butcher, is the best candidate. (paratactic, apposition) His latest book, on singing technique, comes out next week, (hypotactic) Mark is very clever, though erratic. (hypotactic) Frank's cat was found in the coal cellar. (multivariate) Word complex: Mark is a brilliant student. He is very keen. Rover has been lost. Morpheme complex: She is becom mg very nation a/ w/ ic_. Whereas a unit complex involves a grouping of units which can frequently operate as a unit of the rank above, a complex unit is composed of two or more elements typically, but not only, realized by units of the rank below. The relationship between the elements, however, is normally one of dependency involving different grammatical status, e.g. main-subordinate, headword-modifier, root-affix, though it may possibly be one of interdependency. Complex sentence: After Keith was awarded his degree, his confidence increased enormously. Complex phrase: The latest news about Jill Complex word: ab norm al ly hit and miss dyed-in-the-wool In the terms outlined above, the relationships between the basic syntactic units can, then, be summarized by saying that (a) a textual sentence may be composed of a clause or a complex of clauses, a clause complex; (b) a clause contains at least one phrase and more normally a (multivariate) phrase complex; (c) a phrase consists of a single (head)word or of a word complex; and (d) a word may comprise a morpheme or a morpheme complex.

Part II

Class

In systemic grammar members of a unit are grouped together and assigned to a particular class firstly according to the way they function in the next largest unit, secondly according to how they combine with units of the same rank, and thirdly according to the similarities and differences of their internal structure. I consider here the various classes of formal unit.

This page intentionally left blank

3 Word class

Words are traditionally allocated to one of the following range of word classes: noun, pronoun, article, verb, adjective, adverb, preposition, conjunction and interjection; and, anticipating later discussion on the genitive phrase (see Chapter 4, Section 4.7), at the end we add the 'genitive'. Some grammars, e.g. Leech et al. 1982; Halliday 1985a/1994a; Greenbaum and Quirk 1990; Greenbaum 1996; Quirk et al. 1985, include the determiner among the list of word classes. However, we will argue that the determiner is not a separate word class but rather an element of phrase structure which is most typically realized by pronominal subclasses. The allocation of words to word classes is undertaken on the basis of grammatical behaviour. By grammatical behaviour is meant, firstly, the wording environment in which the word appears, its location in the word string and the other words with which it can co-occur; and, secondly, the range of different forms which the word can display. Thus, for example, the word orange can occur in a variety of word contexts and hence belongs to more than one word class. When it denotes a colour, it occurs in contexts such as The (very) orange bulb has blown and The bulb which has blown is (very) orange. With the meaning here orange can be assigned to the word class to which white, pink and new belong. On the other hand, the word orange referring to a fruit occurs in contexts such as This ripe orange is tasty, These ripe oranges are tasty, This fruit is an orange, These fruits are oranges. Orange with this meaning is therefore assigned to a different word class, to which apple, nectarine, pear and subsequently tree, bird, car, house, etc. also belong. Let us take another example. With one of its meanings the word like can occur in the contexts / like peaches, HeI she likes peaches, I I heI she liked peaches. We note that neither of the occurrences of the word orange above has these contexts or variations of form. One does not hear *Iorange ..., * He/she oranges ..., * I/he/she oranged The word like here consequently requires to be assigned to yet another class, to which eat, enjoy, buy and wash belong.

3.1 Noun With respect to the meanings which they convey, nouns denote what we will call 'entities'. Nouns will thus be regarded as a form, indeed the main but not the only form, of nominal word. They may be 'concrete' or 'abstract'. If concrete, the entities referred to are made of material/ physical substance and are thus 'animate' (human or animal), e.g. girl, gorilla, or 'inanimate' (solid or liquid), e.g. chair, water. If abstract, they are intangible, e.g. beauty, depth. Nouns are usually associated with the ability to inflect (i.e. change their form) for the plural, involving either the addition of a syllable to the end of the word or the modification of a word root in some way, e.g. grape - grapes; peach - peaches; mouse - mice; goose - geese; or perhaps a combination of both, e.g. knife - knives; index - indices. In some instances, though, the plural form shows no change from that of the singular, e.g. sheep - sheep; deer - deer. However, the ability to have a plural form does not apply to all nouns. Indeed, as we shall see, these examples above illustrate just one particular subclass of noun known as 'common, count' nouns. Nouns are also associated with the property of being able to follow the articles a and the as well as quantifiers such as some, many, e.g. the car, some warmth. Whilst this property does not apply to all nouns, and most nouns can only follow certain of these words, it is a distinguishing feature of nouns inasmuch as articles and quantifiers do not co-occur in such a relationship with other word classes, e.g. adjective - beautiful, preposition - before, verb - write.

32

Class

Nouns are traditionally divided into the classes proper and common. Proper nouns denote the name of entities which have unique reference. They refer to instances where the entity is viewed as having a single specific or generalized denotation and this includes the names of people, places and geographical locations, days and months, festival occasions, newspapers, magazines and journals, organizations, and institutions, e.g. Ruth, Scotland, Glasgow, Queen Street, the Clyde, Skye, Ben Lomond, Monday, May, Christmas, the Radio Times, the Beano, The Guardian, The Listener, the Scottish Office, the House of Commons, the Scout Association, the University of Strathclyde, the Bank of Scotland. Proper nouns may well be associated in people's minds with the absence of an article (a, the), e.g. Ruth, Mr Bishop, Scotland, Edinburgh, Saturday, May, etc. But whilst this is frequently true, it is not necessarily the case. We do, for example, refer to The Times, the Clyde, the Cairngorms, the Scottish Office, the Hook of Holland. A more significant feature of proper nouns is that they do not, whilst retaining their unique reference, display a variation of article contrast or of plurality. Thus, the examples above at the beginning of the paragraph do not appear as *a Ruth, *the Ruth, * Ruths or *a Scotland, *the Scotland, *Scotlands, e.g. */ saw a Ruth today; * We visited the Scotland this summer. Apparent contradictions of this are to be found in the sentences /'// have a Times, please and We're out of Guardians. However, in such instances the focus has now shifted away from the uniqueness of reference towards the concept of an entity with a multiplicity of possible specimens. Thus a Times denotes one sample from the overall stock of newspapers and Guardians here refers to all immediate specimens - the shop's whole stock - of the paper. The principle of selection from or generality across a multiplicity of instances can similarly be seen in He came on a Thursday and in Mondays can be miserable. These are, therefore, examples of where a proper noun is being treated rather as a countable type of common noun. Common nouns refer to entities which do not have unique reference. They are divided into two classes: count/countable and mass nouns. Count nouns denote entities which can be counted / are countable using ordinary numbers such as one, two, three, four, five, etc., e.g. compass, map, torch, cagoule. Mass nouns are nouns which cannot be counted; here the entity is seen as a mass which cannot be increased or decreased in terms of countable units, e.g. warmth, happiness, furniture, rubbish, milk, enlightenment. So, for example, if milk is added to milk, we merely have more milk, not two milks; similarly, if some milk is poured from a jug into a cup, the quantity of liquid remaining in the jug is less milk rather than, say, threequarters of a milk. The following table compares the behaviour of common count and mass nouns, showing how they combine with the articles a and the and the quantifiers some and many: common nouns count sing, minus art. sing, plus the sing, plus a sing, plus some plural plus many

*car the car a car *some car many cars

mass warmth the warmth *a warmth some warmth *many warmths

A number of common nouns, however, can occur as either count or mass nouns, as illustrated in the (a) and (b) examples below: (a) A dozen eggs, please. (b) You've got egg on your tie. (a) Two teas, please. (b) Tea has gone up enormously.

Word class 33 (a) Look at these fascinating rocks. (b) It's made of solid rock. (a) /'// take three of those cakes, please. (b) Would you like some/ a piece of cake? When nouns can operate in a mixed capacity like this, they are not subject to the constraints which count and mass nouns experience separately. Thus, for example, in different contexts we can refer to cake, the cake, a cake, some cake, many cakes. As we shall see in Part HI, Structure, nouns typically act as the headword element of a nominal phrase.

3.2 Verb In meaning terms, verbs may be said to express processes which can be classified in one of three broad ways. Firstly they can denote actions, e.g. walk, draw, watch, work, feel (pulse), sound (horn), inasmuch as they involve someone doing something and may be identified by questions of the type 'What is X doing / did X do?'. Secondly, they can record events, e.g. occur, collapse, melt, become, see, which involve something happening or a change of state and are identified by 'What is happening / happened?'. Lastly, they can refer to states (a state of affairs, state of mind ...), e.g. be, seem, like, feel (ill), sound (noisy), which relate to any point in time - past, present or future - and which are identified by 'What is / was / will be the state of the subject?'. Systemic grammarians have also classified processes into material, e.g. walk, draw, occur, collapse, melt, mental, e.g. watch, see, listen, hear, expect, like, and relational, e.g. be, seem, become. The grammatical definition of verbs, however, relates to the fact that their form can potentially be inflected/modified to mark tense, aspect, voice, mood and the person form of the subject with which they agree. The variations of the verb work below show some of these inflections in action: tense present: past:

work, works worked

The sequence will work, however, is nowadays generally interpreted not as a future tense - the main verb work has not been inflected - but rather as a marking of future time with the help of the auxiliary verb will. aspect

progressive/continuous: is working, was working perfect: has worked, had worked In these terms is working is the present tense of the progressive/continuous aspect, had worked is the past tense of the perfect aspect, and has been working is the present tense of the perfect continuous aspect. voice

active: passive:

worked, is working was worked, is being worked

34

Class

mood indicative: She works/worked hard. imperative: Work hardier! subjunctive: They insist that she work harder. subject agreement I/you/we/they work; he/she/it works. In practice, the verb in modern English offers little scope for marking agreement with a change of subject. Thus, for example, in the past tense the verb work shows no variation to mark agreement with the subject: I Iyoul he I sheI it I we I they worked. In English it is the present tense of the verb to be which displays the greatest range of subject agreement forms: lam; you/we/they are; he/she/it is. Verbs have been traditionally classified as main or auxiliary. A main verb is one which can operate as the headword or pivotal element of the verbal phrase, e.g. works, worked, is working, will have been worked. If the verbal phrase contains more than one verb word, then the main verb is the rightmost element. There is only a limited range of auxiliary verbs, e.g. do, be, have, will, shall, may, can, must (together with would, should, might, could). As seen above, auxiliary verbs typically co-occur with the main verb, though in conversation it is very possible for repeated mention of the main verb to be omitted, e.g. (Mother) Are you working? (Young daughter, indignantly)

Yes, I am

.

In a tensed verbal phrase, it is the leftmost verb word which carries the tense inflection. That means that in a single word verbal phrase this is the main verb, e.g. works, whereas in a multiple word verbal phrase it is the leftmost auxiliary, e.g. has been working. There are also constraints on the sequence in which the auxiliary verbs can appear, thus will have been being worked is normal, *been have being will worked is not. Unlike main verbs, the first auxiliary verb can also frequently take a contracted and 'clitic' form of the negative n 't, e.g. aren't, shan't, won't, mustn't, can't, mightn't, don't (but not, in standard British English, *mayn't, *amn't). (A few auxiliaries can also take a clitic form with the preceding subject noun, e.g. I'm, he'll, we'd, you've, but in the case of be and have this ability is not restricted to their use as auxiliaries.) Again, auxiliary verbs can be inverted with the subject to form an interrogative, e.g. Do you take sugar?, Will she paint the kitchen?, whereas main verbs typically cannot, e.g. *Take you sugar?, *Paints she the kitchen?. (But in German, for example, such subject-verb inversion is the norm for this type of question.) A further distinction is usually made between primary auxiliaries do, be, have and modal auxiliaries, e.g. can, will, shall, must, may. With the primary auxiliaries some variation is possible in the present tense form of the verb to mark agreement with the subject person, e.g. / am I you are / she is writing. I have I she has written. I do I she does write. With modal auxiliaries, on the other hand, a change of subject does not prompt a change of verb form, e.g. / must I you must / he must write. I can I you can / he can write.

Word class 35 Only one modal auxiliary occurs in a verbal phrase. If a modal is one of several auxiliaries in the phrase, then it is always the first mentioned one, with the remainder being primary auxiliaries. Thus one does not encounter sentences such as * / must can work or * You will may have finished, nor *She be must working or *He have could finished. In addition, there are some further modal auxiliary verbs - dare, need, ought to and used to which are regarded as marginal modals. Dare and need receive this label because whilst on the one hand they can seem to behave exactly as ordinary modal auxiliaries, as in Dare I ask? / / daren 't ask (but note also / didn 't dare ask) and Need we leave so early? / We needn 't leave so early, on the other hand they can sometimes take the infinitive particle to, as in / didn't dare to ask or Do we need to leave so early? / We don't need to leave so early, thus behaving like nonauxiliary verbs. Ought to can take a contracted negative, e.g. You oughtn't to miss the opening, and as part of an ellipted reply the word ought can stand without the infinitive particle, e.g. (Ought I to bring some food?) Yes, you ought. However, in most (though not all) people's speech the infinitive particle is normally required in non-ellipted sentences, e.g. You ought to bring some food rather than * You ought bring some food. Used to expresses the idea that something was the case formerly/ in the past but is no longer, e.g. She used to play very well. (She did play very well can express the same meaning (given appropriate intonation) but, to do so explicitly, should perhaps be followed up with but doesn 't any more.) A number of verbs have a meaning and role akin to that of some auxiliaries above but consist of more than one word and are variously classed as phrasal auxiliaries or semi-auxiliaries, e.g. (must) (will) (can)

have to, have got to, be to, be supposed to, had better, be bound to, be due to; be about to, be going to, be likely to; be able to.

In most instances they have the facility to be able to combine with other auxiliaries and with other semi-auxiliaries, e.g. She must be about to go. He will be able to come. She is supposed to be about to go. He is going to be able to come. The three verbs do, be, have can also operate as main verbs. Compare: They are in France, (main verb) They are living in France, (auxiliary accompanying main verb living) I have an answer, (main verb) I have suggested an answer, (auxiliary verb) She did a good essay, (main verb) She did write a good essay, (auxiliary verb) When used as main verbs, have and be can still be inverted with the subject to form an interrogative, thus: Have you an answer? Are you deaf? But this does not apply to do as a main verb: * Did she a good essay? (In view of this ability of do, be, have to operate as both main and auxiliary verbs together with the fact that the labels auxiliary and main can be thought of as functional positions within the verbal phrase, one current approach (see e.g. Quirk et al. 1985) is to classify verb types not

36

Class

as main verb, primary auxiliary and modal auxiliary but as full verb, primary verb and modal verb. In such a schema, the concepts of main and auxiliary are then handled as structural elements within the verbal phrase.) Verbs display finite and non-finite forms. Finite forms are those which potentially show marking for tense (present or past), mood (indicative, imperative, subjunctive), and agreement with the subject person (first, second or third person): I play, he plays, they play, I played, we played; I am playing, she is play ing, they are playing; I was play ing, they were playing. Play well. Write neatly. It is essential that you be here on time. If I were you,... Non-finite forms are not marked for tense; they occur either as a base infinitive (with or without the particle to): She helped him (to) learn Russian. She watched him learn Russian. She invited him to learn Russian. or as a participle. Although participles are not subject to tense as such, they have two forms, a present and a past participle. They are typically found when the verbal phrase is in a so-called 'compound tense', either where the main verb is non-finite and accompanied by one of the auxiliary verbs be or have (one or more of which may also be non-finite), or when the verbal phrase as a whole is non-finite: Now he is studying Russian. He has studied Russian for two years. He has been studying Russian for two years. Having studied Russian, /found German easy. Russian has been studied here for 40 years.

(present) (past) (past & present) (present & past) (past & past)

Where a participial form is seen as having a noun-like function, it is traditionally called a gerund, e.g. Studying is not difficult; Smoking can damage your health. The label phrasal verb is given to verbs which include after the headword verb element an adverbial particle whose meaning is fused with that of the headword and which is therefore analysed grammatically within the verbal phrase: She switched on the cooker. I looked up the word. He turned off the fire. We found out the cause. There are several differences between phrasal verbs and verbs which are followed by a prepositional phrase. Compare, for example: (i) / looked up the word. (ii) / looked up the chimney. (iii) He turned off the fire. (iv) He turned off the road.

Word class 37 With a phrasal verb it is normally possible to switch the particle to the right of the object. This particle switching is possible in (i) and (iii) above: / looked the word up. He turned the fire off. But in sentences (ii) and (iv), which contain prepositional phrases, this is not possible: */ looked the chimney up. *He turned the road off. Though with phrasal verbs it is perfectly in order to ask questions beginning with the verbal object, of the type What did you look up?, What did he turn off?, and get the replies the word I the fire, it is not possible to pose questions beginning with the adverbial particle, e.g. *Up what did you look?, *Off what did he turn?. With the non-phrasal verbs in (ii) and (iv), however, both forms of question can be posed, the first beginning with the prepositional object, i.e. What did you look up? and What did he turn off?, and the second beginning with the prepositional phrase as a whole, i.e. Up what did you look? and Off what did he turn?. Furthermore the replies can either be confined to the prepositional object or can embrace the whole prepositional phrase, viz. the chimney / up the chimney; the road / off the road. Related to this questioning test is the fact that it is only with non-phrasal verbs that the normal word order can be inverted, thus: Up the chimney I looked. Off the road he turned. but not

* Up the word I looked. * Off the fire he turned. In respect of verbs plus a prepositional phrase it is possible to use the preposition within the relative construction, e.g. The chimney up which I looked... The road off which he turned... The same, however, does not apply to particles within phrasal verbs, thus: *The word up which I looked... * The fire off which he turned... A further point of difference is that in respect of (some) non-phrasal verbs taking a prepositional phrase it is possible to begin the questions with an interrogative adverb in place of a prepositional phrase, e.g. Where did he look?; Where did he turn?. But, related to the fact that the particle in a phrasal verb does not carry its literal, locational meaning, this mode of question is denied to phrasal verbs. In respect of verbs taking a prepositional phrase, it is often possible to modify the preposition, e.g. / looked straight up the chimney. He turned right off the road.

38

Class

This is typically not possible with phrasal verbs with unswitched particles, e.g. */ looked straight up the word. *He turned right off the fire. Some phrasal verbs can in fact be replaced by simple verbs, e.g. switch on - ignite, turn offextinguish, find out - ascertain, hand in - submit, set up - establish. But this is not a requirement nor a distinguishing feature, as may be seen from the fact that He turned off the road (nonphrasal verb) can be replaced by He left the road. Verbs acting as main verbs can be classified as lexical or copular according to whether they have 'content' meaning or merely provide a relational link (denoting the actual/apparent state or change of state) between the subject and the post-verbal completive element (the complement in traditional grammar). lexical verb: Ruth does not sell/ grill/ enjoy/ make cheese. copular verb: Jill isj^seems I appear si lookedjfelt I becameI grew I remained tired. Lexical verbs are traditionally divided into intransitive, transitive or ditransitive according to the number of objects with which they can combine. (See later sections on Complement in Chapters 9 and 10.) Intransitive verbs do not take an object, transitive verbs have one object, and ditransitive verbs have two objects, e.g. intransitive

The ship sank; Keith slept; The liquid cooled; Tom jell; The cats returned home (home here is not an object, it is not 'what' the cats returned but rather 'where'); She wept all night. transitive They sank the ship; I watched the match; First we cooled the liquid; The batsman hti_ the ball with tremendous power; Jill sang an Austrian song. ditransitive They gave her a beautiful watch; We sent him the photos; Pass Jane the butter, please; Mary asked Jill a favour. However, in view of the fact that many verbs can operate in more than one category, cf. Jill sang very well; Jill sang an Austrian song; They gave £10; They gave her £10 it is perhaps more helpful to think in terms of a verb being used intransitively, transitively or ditransitively rather than being assigned to a particular class. Copular or relational verbs combine with a complement - an element which, in the examples below, refers back to the subject:

Word class 39 (a) Ian w the doctor / the best doctor in Glasgow. (identity of the subject) (b) Ian w a doctor / a good doctor. (classification of the subject) (c) Ian is_ clever / rather clever/ very clever indeed. Ian w in good health / in high spirits / out of condition / of no importance. (features/qualities/states of the subject.) 3.3 Adjective Thought of in traditional grammar as a 'describing' word, the adjective has the role of ascribing an attribute or feature to a noun. It may occur either attributively within the nominal phrase, in which case it serves to modify the headword noun, or predicatively outside the nominal phrase to which it relates. Where the adjective functions attributively, then in English it is mainly found before the headword noun, in a pre-head position (prenominal), but in selected contexts it can come after the headword noun, in a post-head position (postnominal). attributive adjective (within the nominal phrase): (a) prenominal: a new car; a responsible child; the principal problem; complete nonsense; (b) postnominal: anything different; somebody new; the child responsible; the court martial; the secretary general; the president elect; the Princess Royal; from time immemorial. Where the adjective functions predicatively, it is mostly in relation to the subject, in which case in English it is separated from the subject by the verb. If, however, the adjective relates to an object nominal phrase, then it will typically be adjacent to the object but structurally distinct from it. predicative adjective (outside the nominal phrase): (a)

This car is new. A child is responsible for this damage. The problem is simple. Jack is fond of Jill.

(b)

We stood the bookcase upright. She drinks her tea black.

Some adjectives can occur in just one of these roles. For example, main, principal, mere and utter are found only attributively in the prenominal position, so one does not hear *The problem is main. Equally, unwell, alone, afraid, asleep, alive and aware normally occur only predicatively, so one does not normally hear *the asleep child. (However, this latter structure may be found where a speaker is giving contrastive focus, e.g. The dead frog .... whereas the alive one ..., or if the adjective is modified by very, e.g. a very aware child.) In English, unlike many other languages, adjectives have a fixed form: their spelling does not vary according to whether the headword noun is singular or plural, or whether it refers to

40 Class a male or a female being. However, some are able to be modified for gradability by comparison or degree. Comparison of the adjective implies (and often includes) reference to a second entity or to a previous state of the first entity, e.g. My bike is cleaner than yours; My bike is cleaner than it was. It is formed either by inflecting the base/absolute form of the adjective (mainly by adding -er and -esi) or by the addition of the words more and most. The -er ending and the word more are used to create the comparative form, and -est and the word most to create the superlative: new, newer, newest', happy, happier, happiest; good, better, best (note the change of adjective stem); luxurious, more luxurious, most luxurious; exciting, more exciting, most exciting. As a rule of thumb, adjectives of just one or two syllables, such as new or happy above, will add the -er and -est inflections, whereas adjectives with three or more syllables, such as luxurious or exciting, make use of the separate words more and most. The norm, however, is not universally adhered to, as can be observed in weather forecasters' references to both cloudier and more cloudy. Gradability of the adjective by degree requires the use of modifying adverbs to mark the relative state or intensity of the adjectival feature in question, e.g. This book is very/most/fairly/ quite interesting. It does not imply or require reference to any second entity and to do so would involve the addition of words such as 'by comparison with the one I read yesterday'. The difference between comparison and degree is illustrated in the following examples: Keith was the most interesting speaker. Keith was a most interesting speaker.

(comparison) (degree)

Only in relation to the first of these sentences could one add of them all, and only in respect of the second would one replace most by very and retain the same meaning. 3.4 Adverb Although many adverbs can also be modified for comparison or degree, e.g. (i) tunefully, more tunefully, most tunefully; (ii) very/most/fairly/quite tunefully, like adjectives they too have an invariable base form. Traditionally in grammar, adverbs have been seen as performing a so-called modifying role in relation to verbs. This role is associated with circumstantial adverbs, which are single words marking the circumstances - how, why, when, where - of the verbal process. verbs, e.g.

She sings beautifully/tunefully/clearly; He came yesterday I annually I herej upstairs.

Circumstantial adverbs can in fact also take on what might be thought of as the 'adjectival role' of modifying nouns within the nominal phrase. nouns, e.g.

the match tomorrow; an away fixture; our holiday abroad.

Secondly, however, adverbs operate in relation to adjectives or to other adverbs, where they indicate the degree of the adjectival attribute or adverbial circumstance. adjectives, e.g. She is quite tuneful / very clever / extremely kind / particularly keen / really grateful.

Word class 41 other adverbs, e.g.

He sings very well / quite tunefully / so^ expressively / amazingly clearly / rather delightfully.

Thirdly, adverbs can modify whole nominal and prepositional phrases or the active or passive verb by focusing in on and heightening or tempering the entity/relator/process concerned. They thus indicate the extent or intensity of the phrase and are known as intensifying adverbs or intensifiers. nominal phrases, e.g.

He is rather a nuisance j almost a teenager / only a lad / quite a character;

prepositional phrases, e.g.

verb, e.g.

He ran almost into the house I fairly near the river/ right at the back / just behind you.

She almost fell; I quite forgot; The house has just been painted; The potatoes have only been peeled (not cooked).

Lastly, adverbs can 'modify' the whole (of the rest) of a clause, though the concept of modification here needs to be interpreted in an increasingly liberal way. Indeed, the relationship is no longer one of modification at all but rather one of interrelationship with the rest of the clause or sentence. Three subgroups of adverb can be listed here. In the first group, known in traditional grammar as 'sentence adverbs', are those which have the function of providing a connective link between the preceding clause and the present one. A more contemporary term is conjunctive adverb. clauses, e.g.

Therefore the performance should now be better. However, it doesn't make sense.

In the second group of adverbs relating to the rest of the clause are those which express different facets of the speaker's perspective on the sentence and which have been labelled modal or disjunctive adverbs, e.g. (i) the speaker's assessment of the degree of certainty or doubt surrounding the factual content of the clause: Jill has probably I possibly gone by bus. Apparently I clearly there was some confusion. (ii) the speaker's comment on / reaction towards the clause content: Amazingly I luckily, no-one was hurt. Unfortunately/regrettably, the car failed its MOT. (iii) the speaker's terms of reference for the clause: Briefly/basically, it doesn 't meet our demands. The third group of sentential/interclausal adverbs embraces expressions regularly used in social discourse and can thus be labelled interpersonal adverbs, though strictly speaking some of the examples below involve more than one word. politeness/courtesy adverbs, e.g. Would you pass me the jam, please. Kindly leave the room.

42

Class

continuity markers, e.g. well, well now, now then, right then. greetings and farewells, e.g. hello, hi, goodbye, cheerio. polarity and agreement responses, e.g. yes, no, okay, all right, certainly, sure. approval formulae, e.g. hurrah, cheers. This range of expressions represents, rather, speaker activity in relation to other participants. (In several instances, they could in fact possibly constitute an utterance entirely on their own, without reference to any other words.) The continuity markers act as attention getters, preparing the way for imminent new discourse. The greetings, politeness expressions, responses and approval words are used by a speaker actually to engage in discourse interaction. Under the heading of adverb are also included several types of particle: the adverbial particle in a phrasal verb, the infinitive marker to, and the negative marker not: (i) A phrasal verb particle has its meaning, which is typically no longer circumstantial, fused with that of the main verb, as illustrated by off in She turned off the fire. (ii) An infinitive particle to frequently, but not always, accompanies the infinitive form of the verb. Indeed, in a sentence like / helped John (to) mend his bike the inclusion of the infinitive particle is optional. (iii) The negative particle not is often associated with verbs / verbal phrases, e.g. did not arrive, and in this context can take the contracted form n'/. But the negative can also modify other classes of word and phrase, e.g. Not a single person withheld their support; I would describe her as not very helpful. However, the contrast might be less obvious between examples such as He isn 't very bright, He's not very bright, and He is not very bright. Across these instances different interpretations are possible, depending on the assignment of word stress and the intonation pattern. There could be a denial or negation of the assertion that he is very bright, i.e. where not is seen as linked to is, or a positive assertion that he is rather dim, i.e. where not is linked to (very) bright, or even an assessment of how bright he is, i.e. where not is linked to very. By contrast with phrasal verb particles, the infinitive and the negative particles have basically only a single form available (to and not /n't) and consequently there is a direct relationship between form and function. Adverbs often used to be thought of as words ending in -ly. A large number of them do, e.g. beautifully, tunefully, clearly, amazingly, fairly, basically, but many do not, e.g. yesterday, here, rather, well, quite, almost, just, very, etc. Equally, of course, some words which do end in -ly are not in fact adverbs, e.g. friendly, likely, lively, lovely, manly. 3.5 Preposition Prepositions have the feature of being accompanied, indeed normally followed, by a completive element in the form of a (single or multiple word) phrase or a clause. phrase:

- in bed, from Mary, a/ home, for sure, on time, between us, without delay; - into the lounge, near the river, oj£the record, inside the house, over the edge, to the hills;

clause:

-from what I heard, near where the bus stops, Ay working all morning, after visiting the museum.

Sometimes the preposition itself is separated from the completive element and placed at the far end of the clause:

Word class

43

Which garage do you go to? Which port are we travelling from? As indicated above, most prepositions in English are positioned before the completive element, that is to say they are 'pre-positioned' (hence their name), but in just a few instances they occur after the completive element, e.g. three weeks ago, that fact apart. For such cases the term postposition has been used. Indeed, in order to account for both possibilities under a single heading Huddleston (1984: 337) uses the term adposition. Although all the examples of prepositions given above are single words, i.e. simple prepositions, there are several complex prepositions, which comprise more than one word, e.g. apart from, out of, owing to, together with, except for, because of, instead of, in view of, in addition to, in spite of, on behalf of, etc. In the context of contrasts such as: She went inj outside/past/through the house', She went inf outside jpast j through some grammarians (see, e.g. Radford1) propose the concept of an intransitive preposition. In the same way that verbs have transitive and intransitive uses, e.g. / don't know the answer; I don't know, it is suggested that contrasts such as She went in the house / She went in; We got off the train / We got off; They climbed up the mountain / They climbed up can be interpreted as reflecting a similar transitive-intransitive alternation for prepositions. However, arguing that prepositions require a completive whereas adverbs do not, we here opt for the traditional analysis in which the 'intransitive' form is handled not as an intransitive preposition but as a prepositional adverb. 3.6 Conjunction Traditionally, conjunctions have been seen as grammatical connectors and are classified into two subtypes: coordinating and subordinating conjunctions. Coordinating conjunctions, also called coordinators or linkers, link units of equal grammatical status, primarily clause with clause (main with main, or subordinate with subordinate) and phrase with phrase. Subordinating conjunctions, also called subordinators or binders, introduce subordinate clauses and join or bind clauses of unequal grammatical status, for example a subordinate clause with a main clause, or one subordinate clause with another subordinate clause which is immediately superordinate. Coordinating conjunctions / coordinators have a purely logical connective function and the principal ones are (both ...) and, but, (either ...) or, (neither ...) nor, as illustrated below: clause:

Ruth plays the flute and Jill sings. Will you go by train or_ will you drive down? The daffodils are out but the tulips are only in bud. He thinks that John is ill and that Audrey is away.

phrase:

Jill and David sang a duet. Stephen has washed and dried the dishes. Will you go by train or by car? She writes both quickly and neatly. He is either very clever or_ very lucky. Neither Jim nor Ken can come.

Thus, andean be equated with the mathematical 'plus' concept, or with 'alternatively', and but

44

Class

with the idea of 'plus, at a tangent'. It should be said that some grammarians use the label 'conjunction' to refer only to coordinating conjunctions. Subordinating conjunctions / subordinators are much more numerous and include, for example, after, although, as, because, before, if, since, until, unless, whether, while, when, where, that, so that, in order that, etc. But in addition to their binding function, they introduce one of two further dimensions of meaning. On the one hand, they incorporate a circumstantial element in terms of which they then introduce a particular type of adverbial clause (see Chapter 5, Clause class), as illustrated by the following examples: He declined because he has no money. She left after the vote had been taken. They will be relegated i£ they lose this match. The trains run although our local station is closed. They went when the rain eased off. He raced down the hill in order to catch the bus. (Further subordinators in this subgroup include as, before, since, until, unless, while and so that.) These subordinators thus both introduce the adverbial subordinate clause and mark the content nature of its circumstantial relationship. In the examples above, this relationship includes reason (because), time (after/when), condition (//), concession (although), purpose (in order). Thus, though not labelled as such in the literature, they may thought of as 'adverbializing subordinators' or 'adverbializers'. On the other hand, subordinating conjunctions may introduce a nominal clause (in transformational generative literature also called a complement clause), as illustrated in: The captain confirmed that Ruth would play. I wonder whether the camera still works. He asked if they could go. Here, in addition, they mark the modal type of the nominal subordinate clause. Thus, in these examples the word that marks the start of a declarative subordinate clause, associated with statements, whereas the words whether and // denote an interrogative subordinate clause, associated with questions. They may thus be seen as 'nominalizing subordinators' or 'nominalizers' (or perhaps 'modalizing subordinators' / 'modalizers'). Indeed, in a transformational framework they are known as 'complementizers'. Although both subordinating conjunctions and prepositions may be followed by a subordinate clause, there is traditionally a difference recognized between them in terms of the nature of the clause patterning. The clause following a subordinating conjunction is tensed and, when separated from the conjunction, is patterned like a declarative clause (subject, which is not a wh-word, followed by verb), as typically found in a statement. Frequently, though not always, if taken out of its context, it could stand on its own grammatically as a main clause, e.g. (because) he has no money -» He has no money; (after) the vote had been taken -» The vote had been taken. The clause following a preposition, on the other hand, if separated from the preposition, does not have this patterning and may be non-finite. For example, in the prepositional sequences from what I heard and after writing the cheque, the clauses what I heard and writing the cheque are not structured as declaratives, the second is also non-finite, and neither of them could stand on their own grammatically as a main clause. One might therefore argue that, though formally analysed as clauses, the post-prepositional sequences what I heard and writing the cheque actually function rather more like phrases than as 'full clauses'. Adding a further potential dimension to the issue, some grammarians, on the basis of data such as

Word class

45

She left after she had drunk her coffee. She left after drinking her coffee. The match was cancelled because the weather was bad. The match was cancelled because of the bad weather. have drawn a parallel between prepositions and subordinating conjunctions which introduce adverbial clauses and have proposed that adverbializing subordinators should be reclassified as prepositions. The argument is that the tensed declarative clauses following the subordinating conjunction can be interpreted equally well as a complement element following a preposition. Though the data may appear compelling and the proposal is very appealing, we do not pursue it here on the grounds that whilst in English some subordinators and prepositions, e.g. after, because, before and until (and even a few prepositional adverbs), do indeed have identical forms, if one turns to some other European languages, one finds that their equivalents do not, cf. conjunction preposition adverb

English after after after / afterwards

French apres que apres apres

German nachdem nach nachher

As in the case regarding the identical form of subject and object nouns in English, therefore, we hold to the view that similarity of appearance between conjunction and preposition conceals a more meaningful difference. 3.7 Interjection Interjections are typically described as those words which are used to express the speaker's exclamation or emotional reaction but which have no further lexical content. They include words like oh, wow, aha, ouch, alas, hey, together with expletives such as damn, golly, hell, etc. The actual scope of interjections is not, however, well-defined, but in any case they are generally regarded as a minor class.

3.8 Article The words the and a/an are classed as the definite and indefinite articles respectively, e.g. the title, a video. Articles are used to mark specificity, generalization or universality of reference to the entity denoted by the following noun. For example, the in the sentences The title has been agreed and The strawberries are ripe makes specific reference to a particular noun entity. Normally the entity concerned has been already mentioned or it is assumed that its identity can be determined from the situation. In Did you hear John on the radio? and /'// take the bus, however, the article is making more generalized reference in that the concern is not with a particular radio or bus - any that 'fit the bill' would do. The indefinite article a in the sentences / know a good restaurant and We've got a ticket makes reference to an entity which is presented as a sample, a member of a type or class but whose specific identity is not an issue, and it may be that the entity is being mentioned for the first time. On the other hand, the articles in The pen is mightier than the sword and A flat can be bigger than a house denote universal or generic reference of the noun entity. 3.9 Pronoun The pronoun used to be thought of as a class of word which may be substituted in place of a

46 Class noun. However, this normally applies only where a noun is the sole element in a nominal phrase, e.g. Cars are very expensive. If, however, the nominal phrase consists of more than one word, e.g. These new cars are very expensive, then typically the pronoun replaces the whole nominal phrase, e.g. They are very expensive. One says that this is typically what happens because there is the exception which proves the rule: so, for example, whilst one would not say * These new they are very expensive, the sentence These new ones are very expensive is perfectly acceptable. A pronoun, further, can sometimes replace the whole of a clause, e.g. // is a new car instead of What you need is a new car. We therefore revise the characterization of the pronoun to say that the pronoun may be substituted for a nominal clause, a nominal phrase or, in just some cases, a nominal word, i.e. a noun. Several subclasses of pronoun may be identified within this category: personal, possessive, demonstrative, relative, typic, interrogative, exclamative, reflexive, reciprocal, emphatic, indefinite, substitute, numeral and quantifier. In addition to functioning in an independent headword role within the phrase (see Chapter 11 on Phrase structure), a number of these pronoun subclasses will be seen to have identical or similar word forms - and with identical meanings - which function in an adnominal, preheadword determiner role (see Section 3.10 on Determiner - a functional element), denoting a subset of the noun entities referred to. In this capacity they are dependent on the rest of the phrase containing the headword noun. Personal pronouns specify the first person (speaker and pro-speaker - /, we), second person (addressee - you), or third person (other referent - he, she, it, they) in the singular and plural. They have only the headword role. I/you/he/she/it/we/they laughed. (also meI himlherI usI them) (Generic)

One shouldn 't laugh.

Note that the first person plural form we, except in such instances as the Gilbert and Sullivan opera line Three little maids from school are we, typically does not refer to several speakers but rather involves one person speaking also on behalf of others, as in We normally finish at 9 p.m. There is, of course, also 'the royal we' exemplified in We are not amused, where on the basis of assumed personal status the speaker adopts the plural form to refer solely to himself or herself. Possessive pronouns mark possession by/ belonging to a person or other entity and fulfil determiner and headword roles. As determiners they form part of a phrase with a noun, whereas as headwords they constitute a phrase on their own and are separated by a verb from the noun or adjective to which they relate, e.g. (determiner) my jyour I his I her jour I their house; (headword) That set is mine/yours/his/hers/ours/theirs. Mine are very comfortable. Note also the historical form of the determiner preserved in lines from some hymns: Mine eyes have seen the glory and God be in mine eyes. (As will be outlined in Section 4.7, Genitive phrase, these possessive examples can be replaced by whole genitive nominal phrases, thus: David's house That set is David's. David's are very comfortable.) Demonstrative pronouns have the deictic function of pointing out a thing/entity. They are divided on the basis of proximity, that is to say, whether the entity referred to is presented by the speaker as being 'near' or 'remote', and they too can operate in either a determiner or headword role. Though nowadays they are typically seen as a class of pronoun, with respect to their determiner role they have also been handled as adjectives.

Word class (determiner)

This model is new. That soup was delicious.

(headword)

This is a new model. That was a delicious soup.

47

In addition to marking the proximity of an entity, demonstratives can also refer to 'What I am talking about / referring to', e.g. This is what you should do. You ... That doesn 't make sense. Relative pronouns have the job of binding a relative subordinate clause either to a particular nominal phrase in the main/superordinate clause or to the whole of the main/superordinate clause. In the latter case the relative clause is known as a sentential relative. Relatives perform a dual role in that, on the one hand, they form an integral part of the content of the subordinate clause and, on the other, they refer back to an antecedent element in the main clause (or to the whole of it) and now within the subordinate clause they act as a substitute for it. (determiner)

The man whose leg is in plaster ... I shall be off at 8 a.m., by which time the newsagent will be open.

(headword)

The thing that pleased us... The car which we hired... The person who bought it... The girl with whom I sang ... The girl who I sang with ... He now trains every evening, which will improve his fitness.

(Although people generally think of relative words as being pronouns, they can in fact also be adverbs, as seen below: The time when/that you fell in the stream was hilarious. The reason why/that she does it is to earn some money. The place where we stopped for lunch was ideal.) The words who, what, which, whose (and the related adverbs), which also occur in interrogative and exclamative sentences, are also known as wh-words. Typic pronouns are used to refer to the type/ the sort of entity. They can also be found in an exclamative context, as above. (determiner) (headword)

Such things aren 't real. Do such exist?

Interrogative pronouns are typically used to form questions about an entity, its possession and determination. Separate from their headword and determiner roles, they can be seen as combining, in an interrogative context, with the roles of personal, possessive, demonstrative and typic pronouns. (personal: (personal: (possessive: (possessive:

headword) headword) determiner) headword)

Who did that? What did she do? Whose mug is this? Whose is the green mug?

48

Class

(demonstrative: determiner) (demonstrative: headword) (typic: determiner)

Which mug is Jill's? Which is Jill's mug? What subject are you studying?

(Like relatives, interrogative words also extend to adverbs, but these have only a headword role: How/why/when/where is it destroyed?.) Exclamative pronouns occur in contexts where a speaker is giving vent to personal feelings about / reactions to an entity or situation. (determiner) (headword)

What a nuisance! What nonsense! What! Another essay!

Reflexive pronouns are used where the object refers to the same entity as the subject. They have only a headword role. (For a consideration of 'subject' and 'object', see Chapter 9 on Clause structure.) This may be (a) a 'direct' object - the entity affected or encompassed by the verbal activity, e.g. I've cut myself. (b) an 'indirect' object - the recipient or beneficiary of the verbal activity, e.g. She's bought herself a new skirt.

or (c) a prepositional object, e.g. She's bought a new skirt for herself. He doesn 't know what to do with himself. (also yourself I itself I ourselves/yourselvesl themselves). Reciprocal pronouns mark the fact that the verbal activity is not just monodirectional but rather interactive, bidirectional or even multidirectional, between/among the participants denoted by the subject. Again they have only a headword role and are found in similar positions to reflexive pronouns. Different from reflexives, however, reciprocals are compound structures. They greeted each other / one another like old friends. They brought a present for each other. Emphatic pronouns serve to emphasize/reinforce the role in the message of a participant who has already been mentioned. They thus modify the main headword of the nominal phrase, most typically occurring immediately after the noun/pronoun that is being emphasized. I myself would write to the manufacturer. (also yourself, etc.) Note, however, the possibility of alternative reflexive or emphatic interpretations in the following: She could have played herself. They should have washed themselves. The reason for the potential ambiguity here is that verbs like play and wash can express an object but do not necessarily have to. To draw attention to an emphatic pronoun interpretation, a comma could be placed before the emphatic pronoun.

Word class

49

Indefinite pronouns denote an entity whose identity is not specific and whose determination is not important to the issue. They too have only a headword role. Somebody/anyone might notice it. (every-/any-/some-/ no- + -body/-one/-thing) Substitute pronouns are used to refer to an entity whose identity has been previously mentioned, or is determinable from the context, and where selection is involved. They typically form the headword of a multiple word nominal phrase, as they substitute for the headword noun alone. Which would you like? I'll have this one. Which one? The green one. Numerals can be thought of as exact numeratives. They embrace two types which are known, firstly, as cardinals or plain numbers, e.g. one, five, forty-nine, and secondly, as ordinals, which list the order of the entities specified, e.g. second, fourth, sixty-fifth. Each of these functions equally in a determiner or headword capacity, e.g. (a) cardinal:

The four climbers unpacked their rucksacks. Four passed.

(b) ordinal:

Her second attempt was successful. The second was better. Jill came second.

Quantifiers. By contrast with numerals, quantifiers may be thought of as inexact numeratives. They specify in non-numerical terms how much of an entity or how many entities is/are being referred to. Most, though not all, quantifiers again function equally in a determiner or headword capacity: AIII bo thj many I some I few I no girls finished. All I both the players arrived. All/both/many/some/few/none All/both/many/some/few/none

finished. of the players arrived.

(also several / each / every / any j much / either / neither / any / a few/a lot / a great deal). Numerative words are sometimes divided into quantitatives and ordinatives. Under such a schema, quantitatives specify the quantity being referred to either as a precise number - using cardinal numerals (exact quantitatives) - or in general terms - quantifiers (inexact quantitatives). Ordinatives mark the position of the entity under consideration in the overall order, either in precise terms using ordinal numerals (exact ordinatives) or in relative, non-numerical terms, employing adjectives, e.g. the next, my last (inexact ordinatives).

Pronominal groupings In terms of the roles that pronouns perform, the various subclasses above may perhaps be grouped under three headings:

50

Class

(1) substantive, referring to / denoting an entity: personal, interrogative, relative, reflexive, reciprocal, indefinite, emphatic, substitute. These pronouns answer the question 'Who?' or 'What?'. (2) determinative, pointing out/to an entity: possessive, demonstrative, interrogative, relative, typic, exclamative. With a headword function they answer the questions 'Whose?', 'Which?', 'What kind?' and with a determiner function they specify 'Whose X?', 'Which X?', 'What kind of X?'. (Note that interrogative and relative pronouns straddle both the substantive and determinative subgroupings.) (3) numerative: (i) quantitatives, indicating a number or quantity: cardinal numeral, quantifier. As headwords they answer the questions 'How many?' or 'How much?' and as determiners, 'How many X?' or 'How much X?'. (ii) ordinatives, marking the numerical or relative position in the order: ordinals. They answer the questions 'Which?' or 'Which X?' in sequential positional terms. 3.10 Determiner: a functional element At this point it should be mentioned that the area of language spanned by these independent headword and determinative forms is a fairly 'difficult' area for grammarians, with respect to the terminology used. In essence the situation arises because the words fulfilling a determiner function are nowadays generally not regarded as adjectives any more, even as limiting adjectives. (See, however, the references by Simpson2 and Lyons3 to demonstrative adjectives, indefinite adjectives, interrogative adjectives.) The reasons given for this are that, unlike adjectives, the determinative forms are not gradable for comparison, cf. The dark/er/est room', The more/most attractive room (but not My/*myer/*myest room), or for degree, cf. A fairly/ very dark room (but not * Fairly/very these problems), and several can not occur predicatively after the verb, cf. The room is dark (but not The idea is *my). On the other hand, BurtonRoberts4 argues that, like adjectives, numerals and quantifiers do co-occur with and follow determiners, e.g. those many books, the one mistake, and quantifiers are gradable, e.g. very many books. Equally, many linguists prefer not to call these words pronouns in their determiner role. This view is not universally adopted, however, as evidenced in for example Freeborn (1987: 61), who remarks that 'Some linguists put determiners in a word-class of their own, but except for the and a/an, all the determiners are pronouns' which constitute 'a special group of modifiers of nouns in NPs'. Hudson5 states that determiners are just pronouns that can combine with a common noun as a dependent and he thus dispenses with the determiner as a word class. (Using this argument, he also includes a and the as pronouns, the difference being that in their case the common noun dependent is obligatory.) Against this background it is therefore interesting how many foreign language reference grammars (of French, Italian, Russian and Spanish, in particular) published in Britain in the last ten years still make no mention of determiners at all and continue to handle this area in terms either of limiting adjectives or of pronouns alone. Finally, note Greenbaum's proposal (1996: 163-4) for a single word class 'determinerpronoun'. His argument is that this grouping of words, including articles, can fulfil an independent (headword) pronominal role or a dependent determinative role, using either the same or similar forms. Some exponents can be pronominal only, some can be determinative only, some can function as both.

Word class

51

The situation is further complicated terminologically by the fact that some grammarians (e.g. Quirk et al. 1985) use the term 'determiner' to denote the umbrella word class together with its several subclasses and the term 'determinative' to mark the function which determiner words (as well as certain phrases, such as Mary's in Mary's lamb) fulfil. The determiner word class is thus almost a mirror image of the functional element of structure 'determinative'. Then just to confuse matters, other grammarians (see, for example, Huddleston 1988a; Downing and Locke 1992) opt for the term 'determinative' as the formal word class and 'determiner' as the functional label, which is the complete reverse of the orientation above. The approach here will be to regard 'determiner' as a functional, specifier element of structure in the nominal phrase, which may be expressed by a range of word subclasses (and by some phrases) which occur in an adnominal position. 'Determinative' then is seen not as a formal word class but rather as a loose term referring to the grouping of some pronominal subclasses which in functional structure can fulfil a determiner role. Indeed, after Greenbaum (1996), it is perhaps more appropriate to think of the word class 'pronoun' as an umbrella grouping of 'determinative-pronominal' subclasses. Following this interpretation, we will enter the pronoun label only where the word is in headword position. Conversely, where the word has a determiner role, at this stage the subclass label alone will be entered. 3.11 Labelling the word classes In syntactic analysis the various word classes will be labelled as follows: noun = noun; pron = pronoun; art = article; verb = verb; adj = adjective; adv = adverb; prep = preposition; sub = subordinating conjunction / subordinator; conj = coordinating conjunction / coordinator; inter] = interjection; gen = genitive (see Section 4.7, Genitive phrase). Pronominal subclasses pers = personal; int = interrogative; rel = relative; poss = possessive; dem = demonstrative; refl = reflexive; rec = reciprocal; indef = indefinite; emph = emphatic; subst = substitute; typ = typic; excl = exclamative; num = numeral; quant = quantifier; ord = ordinative. Note that for the purposes of formal analysis (see Chapter 7 on Formal syntactic analysis), two sets of changes will be made with regard to the labelling of verbs and adverbs. verb adverb

-» aux = auxiliary verb; verb = main verb; ->• adv = adverb / adverbial particle, except that inf = infinitive and neg = negative.

These modifications will not apply within the later functional syntactic analysis, where 'aux', 'inf and 'neg' will be handled not as word subclasses but as functional elements of phrase structure.

52 Class Notes 1 A. Radford (1988) Transformational Grammar, Cambridge: Cambridge University Press, pp. 134-7. 2 J.M.Y. Simpson (1982) A Reference Book of Terms in Traditional Grammar for Language Students, University of Glasgow, p. 12. 3 J. Lyons (1977) Semantics, 2 vols, Cambridge: Cambridge University Press, pp. 452-5. 4 N. Burton-Roberts (1997) Analysing Sentences: An Introduction to English Syntax, 2nd edn, London: Longman, p. 161. 5 R. Hudson (1998) English Grammar, London: Routledge, pp. 36-7.

4

Phrase class

Though in traditional grammar a phrase involved a grouping of two or more words, nowadays the phrase is regarded as a grouping of one or more words which focus around a headword element and which together perform the grammatical role which in other circumstances could be expressed by a single word. A phrase thus inherently involves the expansion around a head element. (Certain phrases will be seen also to incorporate a completive element.) If, therefore, a phrase consists of only a single word, that word is the phrasal headword. Where a phrase consists of more than one word, then the phrasal headword is the one which operates as the pivotal word within the word grouping. In meaning terms, what essentially a phrase does is to express a component element of an idea or 'proposition'. Phrases are formally classified according to the class of word which functions as the headword. So, for example, a nominal phrase has a nominal headword, a verbal phrase has a verbal headword, and so on. (By contrast with elements of clause structure, therefore, the formal classification of phrases is not undertaken in terms of their syntactic function.) Traditionally five main classes of phrase are recognized: the nominal, verbal, adjectival, adverbial and prepositional, and together with these we here also include the genitive and subordinator. We have already mentioned that many systemicists prefer the term 'group' to 'phrase'. Halliday (1994a: 180) actually distinguishes between 'group' and 'phrase': 'A phrase is different from a group in that, whereas a group is an expansion of a word, a phrase is a contraction of a clause.' Against this background he recognizes just one class of phrase, the prepositional phrase, which he describes as consisting of 'a preposition plus a nominal group, for example on the burning deck\ in which the preposition is seen as a minor verb, a minor predicator, having the nominal group as its complement (Halliday 1994a: 212). He goes on to say that the internal structure of across the lake is like that of crossing the lake. But whilst Halliday presents the prepositional phrase as an overarching term for a minor (prepositional) predicator and its complement, he employs no similar overarching term (such as, for example, predicate or predicate phrase) for a major (verbal) predicator and its complement, as for example in We found the map. Equally, if he is to recognize special, phrasal (rather than just group) status for the prepositional phrase because the prepositional headword takes a complement as its norm, then would he not similarly need to do so for adjectival word groupings, in which the adjectival headwords can also take a complement, such as / am keen on/fond of choral music (cf. I enjoy /like choral music)! And what about / am an enthusiast/ a lover of choral musicl (Note that, by contrast with Halliday, Fawcett (1974/81; 1997; forthcoming) is happy to retain the same label 'group' for complete prepositional phrase structures.) In the presentation of the different phrase classes which follows, the focal headword of the phrase is underlined.

4.1 Nominal phrase The nominal phrase has a noun or pronoun as its headword. I prefer the term 'nominal' phrase to 'noun' phrase for several reasons. Firstly, the headword may be a pronoun as well as a noun, that is to say it may be a nominal word, e.g.

54

Class

They are on holiday. This is the way. Hers is the blue cagoule. Six have accepted. Many haven't proper footwear. Secondly, the use of'nominal' then marks a commonality of terminology across units of word, phrase and clause rank, and it will be seen further that a nominal phrase shares a commonality of function with a nominal unit of clause rank. Thirdly, the label nominal phrase gives it a commonality with the other core phrase labels, e.g. adjectival, verbal, adverbial, prepositional. Nominal phrases with noun headwords may typically be preceded and modified (preheadword modification or premodification) or determined by an article, a genitive phrase, a pronoun, an adjective (adjectival phrase) or another noun (nominal phrase), and they may be followed and qualified (post-headword modification or postmodification) by a prepositional phrase or subordinate clause, or in certain cases an adjective or nominal phrase, e.g. this Russian course; my most enjoyable climb; her sister's new bicycle; all our recent holidays; a voice from the past; the song that Jill sang; the secretary general; Jones the butcher. Nominal phrases with pronoun headwords are often assumed to be single word phrases and not to include other words, e.g. She is on holiday. This is the way. Hers is the blue cagoule. Jane has hurt herself. But this is not inevitably so, particularly where the headword is one or an indefinite pronoun, e.g. anybody, someone, or where a numeral or quantifier is involved: The one in the green jacket. The wild one. Anyone could have seen it. Something strange. Six have accepted. The Secret Seven. Several hadn 't got proper footwear. The first few^. Too many. Indeed, even the pronoun forms cited earlier can occur with determinative and adverbial limiter words, e.g. We three are keen. You could win all this. Just hers has been torn. She can blame only herself.

Phrase class

55

4.2 Verbal phrase Verbal phrases have a verb headword, which is a main verb. It may be preceded by an infinitive particle to and/or one or more auxiliary verbs. In the case of phrasal verbs, the main verb is followed by an adverbial particle. starts; started; to start; to have started; to have been started; is starting; was starting; will be starting; has been starting; was started; is being started; might have been being started; switch on; put off; turn down; give in; find out. As can be seen from these phrases, the main verb (underlined above) in all verbal phrases appears as the rightmost verb word. (Note, therefore, that the adverbial particle in a phrasal verb, though within the verbal phrase, is itself an adverb.) Where the phrase is finite, it is the leftmost verb word (main or auxiliary) which carries the tense. (It is perhaps important that the term 'verbal phrase' here should not be confused with 'verb phrase1 as found in transformationally orientated grammars. In that tradition the verb phrase is seen as spanning both the verbal grouping and the following complement phrase(s). Here, however, the scope of the verbal phrase is strictly limited to the verbal group.) 4.3 Adjectival phrase Adjectival phrases have an adjective headword. They may be modified (premodified) by a preceding adverb and qualified (postmodified) by a following adverb, prepositional phrase or subordinate clause, e.g. quick; fairly quick; very quick indeed; quicker than me; much quicker than I was. Typically also adjectival complements / completive elements are handled within the adjectival phrase, e.g. fond of animals; keen on music; interested in history. 4.4 Adverbial phrase Adverbial phrases have as their headword an adverb. This headword adverb may also be premodified by an adverb and qualified by a following adverb, prepositional phrase or subordinate clause, e.g. quickly; fairly quickly; quite quickly enough; more quickly than me; more quickly than I did.

56

Class

(It should be noted that the label 'adverbial' is also used to denote an element of structure of the clause: see, for example, Quirk et al. 1985. This usage is taken up further in the chapters relating to clause structure.)

4.5 Prepositional phrase The prepositional phrase comprises a preposition as the headword plus a second, complement or completive element which is integral to the structure of the phrase. This complement element is most typically realized by a nominal phrase. Indeed, Halliday (1994a: 212) states that 'A prepositional phrase consists of a preposition plus a nominal group'. In practice, however, the completive element can also be an adjectival or adverbial phrase, a further prepositional phrase, or even a subordinate clause. In the examples below we label the class of complement which follows the preposition: in the morning; at_ the beginning in briej\for_ sure after tomorrow; untilfairly recently from off the shelf; to by the tree from what Jill said after listening to the news

(nominal phrase) (adjectival phrase) (adverbial phrase) (prepositional phrase) (nominal subordinate clause) (nominal subordinate clause)

There are, however, one or more instances in which the preposition follows rather than precedes the completive element, e.g. two weeks ago, these problems apart, for which cases the term postposition has been coined. (Ago and apart are not regarded as adverbs here as they cannot appear on their own without the preceding completive element. In this respect two weeks ago thus contrasts with two weeks earlier, where earlier can stand on its own and is an adverb. Similarly, note that these problems apart can be used as a variant form of apart from these problems, containing the complex preposition apart from, and that both expressions can be paraphrased as except for these problems. The meaning of apart above, therefore, differs from that in They now live apart, where apart operates as an adverb.) 4.6 Subordinator phrase Subordinating conjunctions serve to introduce subordinate clauses. Halliday himself (1985a/ 1994a) briefly outlines the 'conjunction group', but generally the concept of a unit (group/ phrase) above word has not been recognized for conjunction/subordinator. In Chapter 3 on word class, a distinction was made between subordinators which introduce adverbial subordinate clauses, which we referred to as 'adverbializers', and those which introduce nominal subordinate clauses - 'nominalizers'. Adverbializers, e.g. because, when, after, etc. mark the particular circumstantial nature of the adverbial subordinate clause. Nominalizers - that, if, whether - have no lexical content meaning but serve to mark the modal type of the nominal subordinate clause. Adverbializing subordinators can readily be expanded by modification, e.g. only i f . . . ; even though ...; ever since .. .;just when; merely because ...; immediately after ...; almost until So although many adverbializing subordinators do consist of just a single word, the fact that they are typically able to be modified by limiting adverbs (or nominal phrases) means that they do thereby enter into phrase relationships and, consequently, need to be handled as phrasal constituents, e.g.

Phrase class 57 Frank will come only i[he can be back by 6 p.m. Even though Jill will be there,... Just because you 've lost your ticket, there's no need... I realized the mistake immediately after they had gone. I realized the mis take five minutes after they had gone. Shortly before she left, she gave Tom a letter. There are instances also where a modifier can be placed in front of a nominalizing subordinates (The point is highlighted when the modifier and subordinator are uttered together and emphasis is placed on the subordinator.) John asked only if he could come. Jill queried just whether the team was fit. Merely that the event had been postponed was sufficient.

4.7 Genitive phrase Genitive phrases are perhaps most readily associated with marking possession, e.g. Jim's car, and are also variously known in the literature as possessive phrases or genitive noun phrases. They can, however, convey other meanings, e.g. The workers'strike (by), The students'entrance (for). These phrases may fill the determiner slot in a larger nominal phrase or form their own independent element, and they can typically be substituted by a possessive pronoun, cf. Genitive phrase: David's sister is here. This book is Ruth's.

Possessive pronoun: His sister is here. This book is hers.

(In contrast to the analysis in Leech et al. (1982: 60, 63, 65), however, we would not see a possessive pronoun also as the head of a genitive phrase but rather, where it does occur as a phrasal headword, as the head of a nominal phrase. The pronoun hers in the example This book is hers above is a fixed form, single word. Unlike Ruth's in This book is Ruth's, it is not composed of separable constituents between which expansion may take place. Compare the rather stilted but nevertheless possible This book is Ruth from Glasgow's.) Genitive phrases are formed by adding an apostrophe and -s at the end of the nominal phrase which specifies the possessor, viz. -'s. They thus comprise a nominal phrase together with a genitive morpheme. This genitive ending is a clitic form which cannot stand on its own but must be attached to an adjacent word. Contrary to some people's understanding, however, it is not a case inflection of the noun, as it is attached to the rightmost word of the nominal phrase. That end word may indeed be the actual headword noun of the phrase but, equally possibly, it may be the last word of a qualifying element within the nominal phrase, as shown below: Mary's / my friend's / the college's car; That car is Mary's / my friend's / the college's; The Principal of the University's appointment; The children's department. Quirk et al. (1985: 328) comment that 'The -s ending is not a case ending in the sense which applies to languages such as Latin, Russian and German. It can be more appropriately described as a "postposed enclitic": i.e., its function is parallel to that of a preposition, except that it is placed after the noun phrase.'

58

Class

If the noun in the particular nominal phrase to which the genitive marker is attached is in the plural and ends in -s, the genitive marker comprises the apostrophe only. the cars' engines; the tadpoles' tails; The director of finances' account. (Some Greek names which end in -s are handled the same way as plurals in -s in that they too add just an apostrophe, e.g. Pythagoras' theorem.) Fawcett (1974-76/1981: 36) labels possessive genitive phrases as 'clusters' rather than phrases (groups) because they cannot, he states, fill elements of structure directly. However, in clauses such as These boots are Ruth's, is that not precisely what the possessive genitive phrase Ruth's in fact is doing? 4.8 Labelling the phrase classes NP = nominal phrase; AdjP = adjectival phrase; PrepP = prepositional phrase; GenP = genitive phrase.

VP = verbal phrase; AdvP = adverbial phrase; SubP = subordinator phrase;

Line marking

I = phrase boundary which is not also a clause boundary.

5

Clause class

With respect to their formal grammatical composition, clauses consist of one or more phrases: What a mess! Enter! Out! (one phrase each) She | left. (two phrases) The best time \ is | before sunrise. (three phrases) Professor Plum \ murdered \ Miss Scarlet \ with a rope \ in the bedroom. (five phrases) In meaning terms what a clause does is to express a single idea or proposition. This may involve giving or seeking information, e.g. / have a flask of coffee back in the car; Have you finished with the dictionary?; it may be concerned with requesting or offering services, e.g. Could you lend me a screwdriver, please?; Would you like some of this cake?, or it may express an emotional reaction, e.g. What a fantastic goal! Clauses which include a verbal phrase are called major clauses and those without a verbal phrase are called minor clauses. major clause: Rita has come. The technician explained that he would need a video camera in order to edit the film. Whether you have a map is not important. minor clause: What a mess! Yes please. No dogs! Fresh strawberries for sale. 5.1 Main and subordinate clauses In terms of their relationships of grammatical dependency, clauses are traditionally classed as main or subordinate. Main clause A main clause is one which, whether or not it can stand on its own without the assistance of a subordinate clause, is not dependent grammatically on a higher node (a node is a position in a

60

Class

diagram of the syntactic structure) or superordinate clause. The main clause is itself a top node clause, as illustrated in the underlined sections of the examples below. I'm going to town. I'm going to town when I've finished this job. When I've finished this job I'm going to town. A main clause which is grammatically well-formed is typically able to stand on its own as a simple sentence in its own right. This applies to I'm going to town above, as it does to the various major and/or minor main clauses below. The fares have been reduced. What have you done? Kindly leave the room! What a fascinating toy! How interesting! Where main clauses include a verb and operate as statements, they generally answer questions of the type 'What did X do?', 'What happened?' or 'What is the situation?'. Although not dependent grammatically on a superordinate clause, a main clause may be dependent in terms of contextual sequence on another main clause in the same sentence. 7 Jack fell down Ruth got out of bed,

2 and broke his crown had a shower,

3 and Jill followed. and got dressed.

In these examples, clauses 1, 2 and 3 are all main clauses; they are all top node clauses and formally of equal grammatical status with each other. But, in terms of contextual sequence, 1 takes place before 2, and both 1 and 2 occur prior to 3. To reverse the sequence of clauses would misrepresent the order of events and depict an entirely illogical situation into the bargain. (These clauses also display grammatical sequencing in relying on the fact that the subjects Jack and Ruth are specified in their particular clause 1 and therefore do not require to be, and on grounds of grammatical cohesion are not, repeated in clause 2 or - in the second example clause 3.) It is, of course, also possible for one main clause to be linked to another main clause in just a list sequencing relationship, without there being any contextual sequencing relationship between them, e.g. The fares have been increased We can hire a machine Jill has been given a woolly lamb

and the service is now less frequent. or we can call in a contractor. but she still likes the brown bear.

Thus, in these instances, we could happily reverse their order and not lose any logicality of meaning: The service is now less frequent and the fares have been increased. We can call in a contractor or we can hire a machine. Jill still likes the brown bear but she has been given a woolly lamb. However, irrespective of whether main clauses are linked through a relationship of contextual or list sequencing, with respect to grammatical dependency they are joined or cooordinated on the basis of equal status.

Clause class

61

Subordinate clause Where a main clause is connected to another clause which is not a main clause, the other clause is a subordinate clause dependent upon the main clause, which is then the superordinate clause. In many people's minds, subordinate clauses are typically associated with major clauses containing a verbal phrase, though as we shall see below they can occur also as minor clauses, without a verb. Functionally, the subordinate clause plays an integral or a supplementary role in relation to the main clause. Where the role is an integral one, the subordinate clause provides a constituent element within the meaning structure of the main clause and may readily be interpreted as operating within the structure of the main clause, in the same manner as a phrasal constituent: main subordinate I've discovered why this light doesn 't work. (cf. I've discovered the problem.) main subordinate The captain announced that his team would score 300 runs. (cf. The captain announced his prediction.) main subordinate How we get there is not important. (cf. This issue is not important.) main subordinate The captain declared when his team had scored 300 runs. (cf. The captain declared with a score of 300 runs.) main subordinate If the weather is good tomorrow, we 'II go to Bridlington. (cf. In those circumstances we'II go to Bridlington.) The principle extends equally to non-finite subordinate clauses: main subordinate / consider Jim to be an excellent leader. main subordinate / want you to come home by 5 p.m. main subordinate I'd like the radio to be put in the kitchen.

62 Class In the context of these integral examples and with them the recognition that a main clause may require a subordinate clause in order to stand on its own, a distinction can be made between the overall span of the main clause and the (possibly more limited) domain of its unsubordinated constituents. This unsubordinated part of a main clause is frequently called the matrix clause, and the first three examples above illustrate well the reason for the difference between a matrix and a main clause. For instance, in the third sentence, How we get there is not important, there is no way that the matrix clause is not important could operate as an independent main clause without the involvement of the subordinate clause How we get there. (For How we get there one could substitute a variety of possible subordinate clauses, such as Whether you go by car, What you did, Who you met, Where you go, When you set off. They would all have the same relationship to is not important.) In such cases and indeed generally where an integral subordinate clause is present, a test can be applied to determine which of the clauses is the matrix part of the superordinate, main clause and which of them is the subordinate clause. If, here, we take the part suspected of being the matrix clause and use it as the basis of a question, we can seek to target a question beginning What? at the suspected integral subordinate clause. Thus, for example, in respect of How we get there is not important, we can ask What is not important? and get the sensible reply How we get there. If, however, we tried to frame the question in reverse, in other words if we tried to use How we get there as the basis of a question which we then begin with What?, we could not even form the question *What how we get there?, let alone expect a sensible reply. On this basis we can confirm that is not important is the matrix part of the main clause and that How we get there is an integral, indeed indispensable, subordinate clause. The question principle illustrated here is one which can be applied to determine both matrix clauses and all types of integral subordinate clause. It is a matter of whether the suspected superordinate clause can be used as the basis for a question such that, if an interrogative word of the type What?, Who?, Which?, How?, Why?, When? or Where? is added to that base, it will form a sensible question which can be targeted at the suspected subordinate clause and can receive a sensible reply using the wording of the subordinate clause. For example, in respect of the sentence The captain declared when his team had scored 300 runs above, we can take the captain declared as the basis for our question, add the question word When and pose the question When did the captain declare? (Note that did declare is merely another way of saying declared.) The reply is provided in the wording of the subordinate clause when his team had scored 300 runs. In other instances the subordinate clause plays a supplementary role and its presence merely provides optional extra information or comment on (the whole or part of) the superordinate clause: main subordinate As you probably know, I shall be in Moscow in May. main subordinate / could eat chocolate every day, which would be very bad for me. main subordinate The president, who works very hard, will be re-elected. (Here the subordinate clause who works very hard provides supplementary information within the structure of the nominal phrase beginning The president.) In these instances, as in all supplementary examples, the subordinate clause could be removed without affecting the grammaticality or altering the meaning content of the remaining main clause, e.g. I shall be in Moscow in May; I could eat chocolate every day; The president will be re-elected.

Clause class

63

Subordinate clauses can also occur as minor clauses without a verbal phrase, e.g. / consider Jim an excellent leader. I find this music very difficult. I expect everything in order for this evening. I want you home by 5 p.m. I'd like the radio in the kitchen. In each case here it can be shown that the NP following the verb has no direct relationship by itself with that verb. The relationship which it does have is firstly and directly only with the following phrase(s), e.g. Jim an excellent leader, this music very difficult, etc. It is only secondarily and indirectly, when taken together with the following phrase(s), that the NP referred to has a relationship with the preceding verb. Thus, the meaning of I expect everything in order for this evening does not include the meaning of I expect everything. In support of this analysis, we can paraphrase the meaning and structure of the whole sentence as / expect that everything will be in order for this evening or I expect everything to be in order for this evening, and in these expanded sentences the added finite or non-finite copular verb now provides an explicit pivot for the relationship between the other phrases in the subordinate clause. Then also, the question What do you expect? can elicit for its reply all the words following / expect from each of the three sentence variants, viz. everything in order for this evening / that everything will be in order ... I everything to be in order — Indeed, the ability to parallel sentences containing a contracted minor subordinate clause with ones containing a major clause is quite normal, as shown by the following contrasts: / consider Jim an excellent leader. I consider Jim to be an excellent leader. I consider that Jim is an excellent leader. I find this music very difficult. I find that this music is very difficult. I want you home by 5 p.m. I want you to be home by 5 p.m. I'd like the radio in the kitchen. I'd like the radio to be in the kitchen. In terms of their role in structure, subordinate clauses are classed as nominal, adjectival/ relative or adverbial.

5.2 Nominal clause Nominal clauses (in traditional grammar known as noun clauses) are subordinate clauses which usually play an integral role in relation to the superordinate clause. They are called nominal clauses because they specify a concrete or abstract entity and typically could be interrogated by the word What? (sometimes also Who?) or replaced in the sentence structure by the pronoun it or that (or he/she). In this way they are seen as performing a 'nominal' function. As indicated under Nominal phrase in Section 4.1, the label 'nominal clause' is chosen in preference to 'noun clause' as a term which is more generalized and which captures the fact that the nominal function is fulfilled both by a subordinate clause or a phrase, which may itself be a single word (a noun or pronoun).

64 Class How you do it is your business. (cf. // is your business.) What you need is a computer. When you go is not important. Whoever wrote this must be crazy. I know what the answer is. (cf. / know it.) We guessed where you went. David said that the tomatoes were nearly ripe. We wondered whether the tomatoes were ripe yet. This is what you should do.

(cf. This is it.)

5.3 Adjectival clause Adjectival clauses are also known as relative clauses, though there is one type of relative clause which is better regarded as an adverbial. They are typically (but not necessarily) introduced by a relative word in the form of a pronoun or adverb. These clauses serve to qualify a foregoing (antecedent) noun headword, and in this way they fulfil a role which potentially might be performed by a simple word. Indeed, the relative clause can provide the answer to a question of the type ' Which + antecedent headword?'. The term 'relative' denotes the fact that the clause relates back to the antecedent noun headword in the superordinate clause and is bound to it by the relative word (with or without a preposition), though this relative word can sometimes be omitted. In analysis the relative word is marked according to the class to which it belongs in the subordinate, relative clause, and this is not necessarily the same as the class of the antecedent in the superordinate clause. relative pronoun:

The people who led the march carried huge banners. (Which people?) The problems which we encountered were soon solved. (Which problems?) The law to which you refer has now been changed. The law which you refer to has now been changed. (Which law?) The problems we encountered were soon solved. The law you refer to has now been changed. (In the last two examples the relative word is omitted.) They didn 't arrive till late evening, by which time the shops had closed. (This is a supplementary not an integral relative and as such is not susceptible to the Which? test.) The lady whose son is in China is over there. (Which lady?) (In these last two examples the relative pronoun has a determiner function.) relative adverb:

The building where I used to work has been knocked down. (The antecedent The building is nominal.) (Which building?)

Clause class

65

The time when she cooked spaghetti was the funniest. (The antecedent The time is nominal.) (Which time?) The time she cooked spaghetti was the funniest. (Relative word again omitted.) It should also be noted that most of these relative clauses could be introduced by the relative word that in place of who or which. This facility does not, however, apply where the relative word is either part of a normal order prepositional phrase, e.g. to which (though with the 'fronted' order which ... to it is fine), or a determiner element within a nominal phrase, e.g. whose son. The people that led the march carried huge banners. The problems that we encountered were soon solved. *The law to that you refer has now been changed. The law that you refer to has now been changed. *The lady that son is in China is over there. The time that she cooked spaghetti was the funniest. Replacement for the adverbial relative where requires the use of a prepositional phrase with that as a fronted completive element: The building that I used to work in has been knocked down.

Reduced relative clauses Sometimes the form of the relative clause is reduced in that not only is the relative word omitted but also the verb is in a non-finite participle or infinitive form and therefore without an auxiliary. If the reduced relative clause is participial, the participle can be in the present (ending in -ing) or past (ending in -en/-ed). (Note that a participial form which may happen to end in -ed should not to be confused with the past tense ending -ed). The antecedent noun headword also has to act as the subject of the relative clause, as the omission of the relative word means that there is no separate specification of the subject in the relative clause. The people leading the march carried huge banners. The man beaten into second place had a broken finger. It is therefore not possible in a relative clause both to omit the relative word and to have a solely participial verb if the subject of that clause is different from the antecedent noun. This can be seen if we try to alter some of the examples above. Compare: The problems which we were encountering were soon solved. * The problems we encountering were soon solved. The law (which/that) you refer to has now been changed. * The law you referring to has now been changed. The time (when) she cooked spaghetti was the funniest. *The time she cooking spaghetti was the funniest.

66 Class Reduced relative clauses in which the verb is expressed in the infinitive can be illustrated as follows, e.g. The person to leave last was Jane. The thing (for you) to remember is... The man (for you) to talk to is Eric. The time (for you) to cut the cake is — Here the relative word that has been omitted is not limited to the subject of the infinitive clause, as may be seen from the following paraphrases in which the relative clauses include the relative words but have finite verbs: The person who left last was Jane. The thing which you should remember is ... The man who you should talk to is Eric. The time at which / when you should cut the cake is —

Restrictive and non-restrictive relatives Except for They didn 't arrive till late evening, by which time the shops had closed, all the examples of relative clauses so far are restrictive or defining in that the relative subordinate clause specifies the nature of the foregoing noun headword. The relative clause in such instances thus integrates with the antecedent headword to form a single unit of functional meaning in which it - the relative clause - serves to define the reference of the nominal phrase. Some relative clauses, however, are optional, extra clauses. In other words they are supplementary clauses, which do not define the antecedent noun headword but provide additional information to it or a comment on it. They are thus non-restrictive or non-defining. John Brown, who is also a lawyer, would make a good candidate. The Lake District mountains, which are quite close to Manchester, are becoming very popular for day trips. (Contrast this with the restrictive/defining relative clause The mountains which are quite close to Manchester are becoming very popular for day trips.) In 1965, when there was no VA T, petrol cost just 25p per gallon. In Austria, where mountain huts are very plentiful, you can do some excellent walking tours. With non-restrictive relative clauses it is normally possible to insert words such as in fact, incidentally, by the way or / might add, thereby emphasizing their supplementary, optional nature. It is also the case that non-restrictive relative clauses neither omit the relative word nor begin with the relative word that.

5.4 Adverbial clause Subordinate clauses which fulfil an integral adverbial function mark the circumstances surrounding the main or superordinate clause. They can specify time (when?), place (where?), reason (why?), purpose (why / what for / to what end?), result (what was the consequence/ outcome?), condition (in what circumstances?), concession (in spite of what?), manner (how?), degree (to what extent?) or preference (instead of what?).

Clause class

67

We 'II set off when Ruth arrives home. They put the jars where no-one could find them. The match was cancelled because it had snowed all morning. She quickly mended her coat in order that / so that she could wear it. He ran so well that he was selected for the county team. If you are going away next week, would you let me know? Although the quality is better, there is still room for improvement. You behave as if/ as though it wasn't your fault. He is doing as well as can be expected. Rather than spend the day indoors, I would sooner work in the garden. Questions to determine integral adverbial clauses are thus formed by adding one of the circumstantial interrogative words/phrases above to the superordinate clauses and targeting it at the suspected subordinate clause, e.g. When will we set off? Adverbial clauses can also fulfil an optional, supplementary role in relation to the main clause, in that they express a connective link to, add a comment on or provide an information sequel to it. What is worse, we didn 't even know. As you all know, we haven't had the official reply yet. We still haven't had an official reply, which doesn 't help matters. There may be a rail strike, in which case I'll go by car. The speaker rose, whereupon the hall immediately went silent. The adverbial clauses in the last three examples represent in fact a type of relative clause which traditionally has been known as a sentential relative. The relative words which and whereupon here do not just link to an antecedent noun but refer back to the whole of the antecedent main clause. Again, similarly to the situation with non-restrictive adjectival clauses, with supplementary adverbial clauses it is normally possible to insert words such as in fact, incidentally, by the way or / might add. It has been argued that sentential relative clauses are not subordinate clauses at all but rather main clauses. The reasons offered for this claim are twofold. Firstly, in some texts sentential relative clauses can be found as separate sentences, in which the relative wh- word begins with a capital letter, e.g. We still haven't had an official reply. Which doesn't help matters. Secondly, it is explained that the relative wh- word can be replaced by and this, thus: We still haven't had an official reply, which doesn't help matters. We still haven't had an official reply, and this doesn't help matters. There may be a rail strike, in which case I'll go by car. There may be a rail strike, and in this case I'll go by car. The speaker rose, whereupon the hall immediately went silent. The speaker rose, and upon this the hall immediately went silent. Against these points, however, I would counter, firstly, that any occurrence of a sentential relative clause as a textual sentence is purely an orthographical feature of the writer's style, designed to highlight the relative clause. But that does not alter the subordinate status of the clause. It rather points to the possibilities of interplay between textual form and grammatical status for effect. Secondly, I would point out that, whilst the wh- word can indeed be replaced

68

Class

by and this, the sentential relative clause in fact fails the ordinary, standard coordination test. One cannot successfully add just the conjunction and to link the main clause with the unaltered sentential relative: We still haven't had an official reply, which doesn 't help matters. * We still haven't had an official reply, and which doesn't help matters. There may be a rail strike, in which case I'll go by car. * There may be a rail strike, and in which case I'll go by car. The speaker rose, whereupon the hall immediately went silent. *The speaker rose, and upon which the hall immediately went silent. 5.5 Labelling the clause classes main clause = main clause; sub.clause = subordinate clause. Line marking 11 = clause boundary which is not also a sentence boundary.

6

Sentence class

6.1 Formal types of sentence The formal classification of sentences is traditionally in grammatical studies undertaken according to the number and class (main or subordinate) of clauses they contain. Where a sentence consists of a main clause only, it is known as a simple sentence.

simple sentence: I'm off to town now. Did you get a newspaper? Do tell me the story of the red monk. What a price these cagoules are! Where the sentence contains two or more main clauses, it is known as a compound sentence.

compound sentence: She'splaying hockey \\ and then she's going to the theatre. I've tried all day \ \ but I still can't get through. Would you like chicken \\ or do you prefer beef? A sentence which comprises a main clause and at least one subordinate clause is called a complex sentence.

complex sentence: When we've washed up \\ we'll have a game of dominoes. If you don't like the heat \\ get out of the kitchen. She bought three bars \ \ because she adores the taste. We went via Birmingham \ \ in order to save time. What he thinks \\ is not important. That is 11 what was announced. Has Jill said \ \ whether she 'II be available? This is the draft \ \ which I wrote yesterday. The line is to be closed \ \ which doesn 't surprise me. Lastly, where a sentence consists of at least two main clauses and at least one subordinate clause, it may be termed a compound-complex sentence.

70

Class

compound-complex sentence: If you need a hand \ \ give me a call \ \ and I'll pop round. Our plans are now complete \ \ and an order will be placed \ \ as soon as we have received the quotations.

It has to be said, however, that this formal classification of sentences is rather arbitrary and as such not particularly helpful. Thus, for example, the compound class includes sentences with any number of main clauses, provided it is more than one. The complex label does not place any constraint on the number of subordinate clauses which may be present, nor does it provide a way of recording dependency relationships which may exist between two or more subordinate clauses in a sentence. Analysis of a text based solely on a classification of this sort would therefore give only a partial indication of the degree of the complexity of its sentence structure.

6.2 The sentence as clause complex Grammatically, as indicated above, a sentence consists of one or more clauses which are interrelated on a coordinate or subordinate basis. Given that, in terms of meaning, each clause expresses an idea or proposition, then what the sentence does is to express one or more ideas or propositions, interwoven to present a coherent whole. (Arguments used to abound regarding the notion of completeness of the ideas. But irrespective of how this concept might best be defined, we take the view that it is not a productive line to pursue and that it is not necessary to think either of each idea or of the totality of ideas as being 'complete'. There is surely always another relevant detail which could have been added.) In these terms the sentence can be seen as a unit of textual structure which combines and interacts with other sentences to form larger coherent units of text, such as paragraph and chapter. By 'coherent' is meant the logical sequencing and interrelatedness of ideas. Of course, in order to help ensure that the coherence of the message content reads smoothly, a writer would normally employ the various grammatical devices of cohesion, as mentioned under 'textual function' in Chapter 1, Section 1.3, Language functions and the semantic stratum. Some linguists indeed choose to reserve the label 'sentence' for the unit of textual study. In such circumstances, the preferred term when discussing units of grammatical form would be clause complex. As outlined in Chapter 2 (Grammatical framework), this relates to the logical grouping of clause units around at least one main clause and in any configuration. The scope of the textual sentence (bounded by an initial capital letter and a final full stop) and the grammatical clause complex might normally be expected to be the same. For example, the strings This semester there will be a course on formal syntax. Then next semester there will be a module on functional syntax, are presented as two sentences and two (simple) clause complexes. They could easily be coordinated into the single sentence and a single (now compound) clause complex This semester there will be a course on formal syntax and then next semester there will be a module on functional syntax. However, it is quite possible for the scope of sentence and clause complex to diverge. Thus, whilst the sequence Ruth has passed her violin exam, which should please her parents is one sentence containing one clause complex, it could for effect be written as two sentences Ruth has passed her violin exam. Which should please her parents. But the two sentences would still involve only the one clause complex, because the second sentence - a sentential relative - remains a subordinate clause. I adopt a middle way. Although I accept that the sentence is primarily a textual unit, I also see it as one which by tradition in linguistic description serves as a unit of grammatical form for the purpose of analysing grammatical structure. In this capacity and within the standard orthographical boundaries, the sentence is a cohesive grouping of one or more clauses, i.e. a clause complex.

Sentence class 71 6.3 Labelling the sentence #S# = Sentence Line marking III = sentence boundary.

7

Formal syntactic analysis

Systemic grammar was not originally developed for formal syntactic analysis. The grammar is orientated rather towards functional syntax, where reference to formal units and classes is made in order to mark how a functional element is realized. In a functional syntactic analysis the roles of the elements of structure are shown, and each structural element is set out on an equal basis. The framework thus does not need to use a hierarchical tree patterning for marking dependency; instead it employs a rank-based mode of analysis which leads to an essentially flat tree patterning. If, as now, this patterning is applied to the analysis of formal units alone, it will not register, for example, the different relationship which can obtain between NPs and a VP. Thus, if one takes a sentence such as John gave his friend the map, a formal analysis of the clause will record merely 'NP + VP + NP + NP', showing that its structure (and the meaning) involves elements realized by these four classes of phrase. It does not, at this formal stage, mark any particular relationship between them. Nevertheless, in order to illustrate the impact of an analysis in terms of units and classes, we will, subject to the two (temporary) modifications within word class, apply the framework which has been built up to a range of example sentences. The description will mark off sentence units and analyse sentences in terms of their clause constituents, specifying the dependency class of each clause. It will then analyse clauses in terms of their phrase constituents, stating the class of phrase. Finally, it will analyse phrases in terms of their word constituents, indicating the class of word. Not infrequently, of course, a clause has to be analysed as operating within the structure of a clause, or a clause or phrase within the structure of a phrase. Line marking

||| = sentence boundary 11 = clause boundary which is not also a sentence boundary I = phrase boundary which is not also a clause boundary Labelling

Sentence

#S#

Clause

main clause = main clause;

sub.clause = subordinate clause.

Phrase

NP = nominal phrase; AdjP = adjectival phrase; PrepP = prepositional phrase; GenP = genitive phrase.

VP = verbal phrase; AdvP = adverbial phrase; SubP = subordinator phrase;

Word

noun = noun; pron = pronoun; art = article; verb = verb; adj = adjective; adv = adverb; prep = preposition; sub = subordinating conjunction / subordinator; conj = coordinating conjunction / coordinator; interj = interjection; gen = genitive.

Formal syntactic analysis

73

Pronominal subclasses pers = personal; int = interrogative; rel = relative; poss = possessive; dem = demonstrative; refl = reflexive; rec = reciprocal; indef = indefinite; emph = emphatic; subst = substitute; typ = typic; excl = exclamative; num - numeral; quant = quantifier; ord = ordinative. (As mentioned in Chapter 3, Word class, the pronoun label will be entered only where the pronoun is in headword position. So where a pronoun has a determiner role, only the subclass label will be entered.) For the purposes of formal analysis, two temporary sets of changes are made with regard to the labelling of verbs and adverbs: verb adverb

aux = auxiliary verb; verb = main verb; adv = adverb / adverbial particle, except that inf = infinitive and neg = negative.

These modifications will not apply within the later functional syntactic analysis, where 'aux', 'inf and 'neg' will be handled not as word subclasses but as functional elements of phrase structure.

(* The relative pronoun what here would be described as a 'fused' or 'free' relative pronoun, where the antecedent entity to which it refers is fused within the relative word and where the single word what can in consequence be paraphrased as that which.)

74 Class

(* - another fused/free relative)

Formal syntactic analysis

(the butcher here operates in a defining capacity to Jones)

75

76

Class

(a well-known lawyer operates in a supplementary capacity to John Smith)

Formal syntactic analysis

77

78

Class

Formal syntactic analysis

79

This page intentionally left blank

Part III

Structure

Structure accounts for the composition of a unit in terms of its functional elements and for the nature of the relationship which can be established between these elements. In keeping with the view of a functional orientation outlined earlier, I shall here adopt a top-down approach to the presentation of structure, from sentence through to word.

This page intentionally left blank

8

Sentence structure

8.1 Free and bound elements For the description of the functional structure of the textual sentence and in place of the formal labels main and subordinate, initially two basic elements can be recognized: free and bound, as illustrated below. free:

/ have planted the seeds. John explained that the trip was off. After she had been to the meeting, Jill was furious.

bound:

John explained that the trip was off. After she had been to the meeting, Jill was furious.

Like a main clause, a free element of sentence structure needs to be capable of functioning as a simple sentence. In the first sentence / have planted the seeds, which consists only of a main clause, fulfils this requirement. But in the second and third sentences, the matrix parts of the main clauses, John explained and Jill was furious, rely on the subordinated elements to complete the message structure of the sentence. Thus, whilst technically it is true that John explained on its own can make sense, the situation being presented in John explained that the trip was off is not so much that John was the person who did the explaining but rather that John explained something, and it is the bound clause which expresses that something. The bound clause thus acts as an element which is completing the sense of the sentence. Likewise, //// was furious would stand grammatically on its own, but in the situation being presented here she was furious in the circumstances of having been to the meeting, and again it is the bound element which provides the information about those circumstances. In support of this interpretation we can point to the fact that in each sentence the subordinate clauses can be replaced by phrases, e.g. John explained the situation; After the meeting Jill was furious, and that these replacement phrases would be automatically analysed as elements within the structure of the main clause. In the context of the meaning of each of these two sentences, therefore, the scope of the free element extends across the whole sentence and thus includes the bound element. The bound clause is thus playing an integral role within the free clause. Subordinate clauses which have this type of role will respond to the question words and phrases Who?, What?, Which?, How?, Why?, When?, Where?, In what circumstances?, etc., which when added to the superordinate clause may be targeted at the bound clause. I should perhaps comment that, although the terms 'free' and 'bound' strictly relate to elements of structure whilst 'main' and 'subordinate' refer to formal classes of clause, in practice one readily encounters the terms 'free clause' and 'bound clause'. If, however, we consider a sentence such as There may be a rail strike, in which case I will go by car, consisting also of a free and a bound element, then we find a different relationship in operation. As was said above, bound elements play a role in relation to the free element. However, in terms of the message structure here the bound element in which case I will go by car does not supply a completive element or circumstantial constraint to the free element but rather serves to add an optional, supplementary comment. The same is true in sentences such as The tax on petrol has been increased, which won't please Dave or even The tax on petrol, which has never been low, has been increased again.

84 Structure Yet, although a classification of elements of sentence structure into free and bound may on the basis of these few examples seem adequate, it is unable to take account of further degrees of grammatical dependence between bound elements. For example, in the sentence John knew that we would be leaving before the meeting had finished, it is very indelicate, that is to say lacking in detail, merely to say that the whole constitutes a free element within which that we would be leaving and before the meeting had finished are bound elements. Using the question procedure outlined above (here forming the question 'What did John know?'), we can first establish that that we would be leaving is directly dependent on and bound to (the unsubordinated, matrix part of) the free element. However, before the meeting had finished is bound immediately, not to the matrix part of the free element, but to the first bound element, that we would be leaving', it answers the question 'When would we be leaving?'. In other words, the before clause is dependent on the free element only indirectly and through the that clause. To record this greater degree of detail, or delicacy as it is referred to, the analysis of elements of a sentence is undertaken using the letters of the Greek alphabet: alpha (a), beta (0), gamma (y), delta (5), epsilon (e), zeta (g), eta (17), theta (0)..., where a represents a free element and /?, y, etc. mark the descending degree or immediacy of dependency of bound elements on the alpha. (Initially here, for reasons of simplicity of representation, the elements are labelled in a flat linear mode. This will be modified later.) a ft y \\\John knew \\ that we would be leaving \\ before the meeting hadfinished.\\\

Let us further illustrate the procedure with a more complex example: a ft y I1 \Jirn said \ \ that Bill had stated \ \ that he couldn 'tgo\\ 5 e if advance payment was required \ | before he got his salary. \ \ \

In this second example we establish (a) that that Bill had stated is immediately bound to Jim said (by being able to frame and answer the question 'What did Jim say?'), (b) that that he couldn't go is bound to Bill had stated (by posing the question 'What had Bill stated?'), (c) that if advance payment was required is dependent on he couldn't go (with the question 'In what circumstances / why couldn't Bill go?'), and (d) that before he got his salary is bound to if advance payment was required (using the question 'When was advance payment required?'). In this way the analysis is able to show the degree of depth or distance in grammatical relationships between the various bound elements and the (matrix part of the) free element. Thus, in this last example, the epsilon element before he got his salary is dependent on the alpha element only through the intermediary relationship provided by the delta, gamma and beta elements: in other words the epsilon element stands at four degrees of grammatical depth removed from the matrix part of the free, alpha element. Indeed the depth relationship between bound clauses is particularly well illustrated by the nursery rhyme This is the house that Jack built. a 0 y I11 This is the maiden all forlorn \ \ that milked the cow\\ that tossed the dog d

B

C

\\that worried the cat \\ that killed the rat \\ that ate the malt \\ r, e that lay in the house \ \ that Jack built. \ \ \

Sentence structure

85

It is, of course, possible for the same bound clause to stand in a different dependency relationship to the alpha clause, depending on the nature and structure of the total sentence; this will, inevitably, reflect the difference in meaning structure between sentences. The point can be simply illustrated by contrasting the analysis of the first example above with that of a sentence which still involves the same three clauses but in an alternative patterning and reflecting a different meaning. ft y a \\\John knew \\ that we would be leaving \\ before the meeting hadfinished.\\\ ft a ft I I I Before the meeting had finished \\ John knew \\ that we would be leaving.\\\

(Strictly speaking, this analysis requires some modification, if it is to show that John knew and that we would be leaving have a primary bond and then together relate to the clause Before the meeting had finished, e.g. ft aa a/3 \\\Before the meeting had finished \\ John knew \\ that we would be leaving. \\\ See Section 14.1.1, Linear and embedded recursion.) We can do a similar restructuring exercise with the second, more complex sentence: « ft y 11 \Jirn said \ \ that Bill had stated \ \ that he couldn 't go\ \

d E if advance payment was required \ \ before he got his salary. \ \ \ a ft ft 11 \Jirn said \ \ before he got his salary \ \ that Bill had stated

y

o

\\that he couldn't go \\ if advance payment was required.\\\ a ft y 11 \Jim said \ \ that Bill had stated \ \ before he got his salary y 6 \\that he couldn't go \\ if advance payment was required.\\\

In these (last two) restructured sentences, the relocation of the clause before he got his salary changes the meaning such that it now answers, on the one hand, 'When did Jim say ...?' and, on the other, 'When had Bill stated ...?'. ft a ft 11 \Ifadvance payment was required \ \ Jim said \ \ that Bill had stated

y